{ "schema_version": "1.0", "as_of": "2026-09-15", "organizations": [ "McKinsey", "BCG", "Bain", "PwC", "EY / EY-Parthenon", "KPMG", "Accenture", "Oliver Wyman", "Palantir" ], "classifications": { "S": "Strong evidence", "M": "Moderate evidence", "W": "Weak mention", "N": "No meaningful evidence", "C": "Contradictory evidence" }, "statement_type": "SYNTHESIS", "method": "Editorial coverage of the specific reviewed publications, not a survey, independent validation, organization endorsement or buyer-demand finding. Broad composite themes do not prove all subfunctions. No meaningful evidence means not established in this tranche.", "public_definition_amendments": 6, "rows": [ { "requirement": "workflow_redesign", "label": "Workflow, role and process redesign", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-seizing-the-agentic-ai-advantage", "mckinsey-company-the-agentic-organization-contours-of-the-next-paradigm-for-the-ai-era" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/people%20and%20organizational%20performance/our%20insights/the%20agentic%20organization%20contours%20of%20the%20next%20paradigm%20for%20the%20ai%20era/the-agentic-organization-contours-of-the-next-paradigm-for-the-ai-era.pdf" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-the-widening-ai-value-gap-build-for-the-future-2025", "bcg-from-potential-to-profit-closing-the-ai-impact-gap", "bcg-ceos-are-starting-to-see-value-from-ai-now-comes-execution" ], "source_urls": [ "https://media-publications.bcg.com/The-Widening-AI-Value-Gap-Sept-2025.pdf", "https://www.bcg.com/publications/2025/closing-the-ai-impact-gap", "https://www.bcg.com/publications/2026/how-ceos-scale-ai-value" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-state-of-the-art-of-agentic-ai-transformation", "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why", "bain-company-ai-is-transforming-productivity-but-sales-remains-a-new-frontier" ], "source_urls": [ "https://www.bain.com/insights/state-of-the-art-of-agentic-ai-transformation-technology-report-2025/", "https://www.bain.com/insights/your-ai-budget-is-growing-your-returns-arent-heres-why/", "https://www.bain.com/insights/ai-transforming-productivity-sales-remains-new-frontier-technology-report-2025/" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-want-roi-from-ai-go-for-growth-pwc", "pwc-ai-agent-survey-pwc" ], "source_urls": [ "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-roi.html", "https://www.pwc.com/us/en/tech-effect/ai-analytics/ai-agent-survey.html" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-can-ai-advance-toward-value-if-workforce-tensions-linger-ey-global", "ey-ey-parthenon-how-ai-innovation-powers-microsofts-finance-journey-ey-global", "ey-ey-parthenon-how-microsoft-uses-genai-for-compliance-ey-global" ], "source_urls": [ "https://www.ey.com/en_gl/insights/workforce/work-reimagined-survey", "https://www.ey.com/en_gl/insights/consulting/how-ai-innovation-powers-microsoft-s-finance-journey", "https://www.ey.com/en_gl/insights/law/how-microsoft-streamlined-regulatory-compliance-with-gen-ai" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-kpmg-managed-services-outlook-survey-2026", "kpmg-global-ai-pulse-q1-2026" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/transformation/managed-services-outlook.html", "https://assets.kpmg.com/content/dam/kpmgsites/xx/pdf/2026/04/global-ai-pulse.pdf" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-the-front-runners-guide-to-scaling-ai", "accenture-systemic-ai-at-the-root-of-manufacturing-performance", "accenture-the-new-rules-of-platform-strategy-in-the-age-of-agentic-ai", "accenture-the-ai-enabled-capital-project" ], "source_urls": [ "https://www.accenture.com/en/insights/ai-data/front-runners-guide-scaling-ai", "https://www.accenture.com/en/insights/digital-engineering-manufacturing/ai-manufacturing", "https://www.accenture.com/en/insights/strategy/new-rules-platform-strategy-agentic-ai", "https://www.accenture.com/en/blogs/infrastructure-capital-projects/ai-enhanced-capital-project" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-turning-health-system-ai-spending-into-bottom-line-results", "oliver-wyman-how-agentic-ai-is-reshaping-the-services-industry", "oliver-wyman-why-government-ai-success-starts-with-simpler-operations", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/perspectives/health/2026/mar/turning-health-system-ai-spending-into-bottom-line-results.html", "https://www.oliverwyman.com/our-expertise/insights/2026/aug/how-agentic-ai-reshapes-services-firms.html", "https://www.oliverwyman.com/our-expertise/insights/2026/jul/government-ai-success-starts-with-simpler-operations.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-government-ai-success-starts-with-simpler-operations", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-overview-ontology", "palantir-cleveland-clinic-palantir", "palantir-sompo-and-palantir", "palantir-impact-airbus-and-skywise" ], "source_urls": [ "https://www.palantir.com/docs/foundry/ontology/overview", "https://www.palantir.com/impact/cleveland-clinic/", "https://www.palantir.com/impact/sompo/", "https://www.palantir.com/impact/airbus/" ], "editorial_note": "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-overview-ontology" ] } }, "interpretation_limits": [ "Dominant cross-firm finding; this is business process redesign, not the narrower machine execution-control category.", "Historical operating-model evidence is not new agent adoption.", "Retain only as editorial coverage of workflow redesign. This is not proof of deployment, efficiency gains, an execution-control architecture, or purchase intent." ], "strong_organization_count": 9 }, { "requirement": "data_context_integration", "label": "Data, context and integration", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-agentic-ai-is-rewriting-the-rules-of-data-risk-management", "bcg-the-agentic-leadership-playbook-a-scaling-strategy-for-ctos-and-cios" ], "source_urls": [ "https://www.bcg.com/publications/2026/managing-data-risk-in-the-age-of-agentic-ai", "https://www.bcg.com/publications/2026/agentic-ai-strategy-cio-cto" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-why-agentic-ai-demands-a-new-architecture", "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why", "bain-company-the-architecture-decisions-only-cfos-can-make" ], "source_urls": [ "https://www.bain.com/insights/why-agentic-ai-demands-a-new-architecture/", "https://www.bain.com/insights/your-ai-budget-is-growing-your-returns-arent-heres-why/", "https://www.bain.com/insights/the-architecture-decisions-only-cfos-can-make/" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-how-ai-is-transforming-research-and-development-pwc", "pwc-want-roi-from-ai-go-for-growth-pwc", "pwc-ai-in-operations-pwc" ], "source_urls": [ "https://www.pwc.com/gx/en/1/services/ai/ai-research-development.html", "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-roi.html", "https://www.pwc.com/gx/en/1/services/consulting/technology/ai-operations.html" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-how-ai-innovation-powers-microsofts-finance-journey-ey-global", "ey-ey-parthenon-how-ai-drove-data-optimization-for-oil-and-gas-capital-projects-ey-global", "ey-ey-parthenon-how-novartis-optimized-its-supply-chain-ey-global" ], "source_urls": [ "https://www.ey.com/en_gl/insights/consulting/how-ai-innovation-powers-microsoft-s-finance-journey", "https://www.ey.com/en_gl/insights/energy-resources/how-ai-drove-data-optimization-for-oil-and-gas-capital-projects", "https://www.ey.com/en_gl/alliances/novartis-optimizes-supply-chain-with-real-time-risk-insight" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-kpmg-global-ai-in-finance-report", "kpmg-global-tech-report-2026", "kpmg-kpmg-managed-services-outlook-survey-2026" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/kpmg-global-ai-in-finance-report.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/global-tech-report.html", "https://kpmg.com/xx/en/our-insights/transformation/managed-services-outlook.html" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-building-a-reinvention-ready-digital-core", "accenture-systemic-ai-at-the-root-of-manufacturing-performance", "accenture-data-essentials-with-generative-ai", "accenture-the-ai-advantage-for-insurers" ], "source_urls": [ "https://www.accenture.com/en/insights/technology/building-reinvention-ready-digital-core", "https://www.accenture.com/en/insights/digital-engineering-manufacturing/ai-manufacturing", "https://www.accenture.com/en/insights/ai-data/new-data-essentials", "https://www.accenture.com/en/insights/insurance/ai-advantage-insurers" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-how-ai-will-reshape-the-frontlines-of-upstream-oil-and-gas", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2025/oct/how-ai-oil-gas-redefines-competitive-advantage.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-overview-ontology", "palantir-aip-chatbot-studio-overview" ], "source_urls": [ "https://www.palantir.com/docs/foundry/ontology/overview", "https://www.palantir.com/docs/foundry/chatbot-studio/overview" ], "editorial_note": "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-overview-ontology" ] } }, "interpretation_limits": [ "Semantics, access, freshness and integration are frequent primary bottlenecks.", "Integration may be a more immediate budget than portable proof.", "Retain as editorial coverage of governed data, semantics and integration. Do not infer a complete action contract or treat integration as solved by protocol syntax." ], "strong_organization_count": 9 }, { "requirement": "runtime_policy", "label": "Runtime policy and execution controls", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ], "source_urls": [ "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-what-is-ai-governance", "bain-company-the-three-layers-of-an-agentic-ai-platform" ], "source_urls": [ "https://www.bain.com/insights/what-is-ai-governance/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-how-do-you-win-the-game-of-ai-telephone-pwc", "pwc-who-s-in-control-managing-the-cyber-risks-of-ai-agents", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-telephone.html", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-control", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-case-study-building-an-enterprise-scale-agentic-ai-os-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_gl/insights/ai/building-an-enterprise-scale-agentic-ai-operating-system", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value", "kpmg-kpmg-and-microsoft-scale-trusted-enterprise-ai-agents-globally-through-deployment-of-agent-365-and-copilot" ], "source_urls": [ "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html", "https://kpmg.com/xx/en/media/press-releases/2026/06/kpmg-and-microsoft-scale-trusted-enterprise-ai-agents-globally.html" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ], "source_urls": [ "https://www.accenture.com/en/blogs/security/strengthening-ai-agent-security-identity-management", "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-how-ai-will-reshape-the-frontlines-of-upstream-oil-and-gas" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2025/oct/how-ai-oil-gas-redefines-competitive-advantage.html" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-action-types-submission-criteria", "palantir-action-types-permissions" ], "source_urls": [ "https://www.palantir.com/docs/foundry/action-types/submission-criteria", "https://www.palantir.com/docs/foundry/action-types/permissions" ], "editorial_note": "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market. PAL 03/04 are direct native-product documentation for submission checks; a claim that only Nikxius prevents actions would contradict this evidence.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-submission-criteria", "palantir-action-types-permissions" ] } }, "interpretation_limits": [ "Explicit shared execution or policy layer. Does not establish a separately monetized market.", "Palantir supplies native product evidence, creating a bundling countercase.", "Distinguish recommendations for runtime governance from documented enforcement in a product. The cited publications do not prove implementation quality, unavoidable mediation, or a separate market." ], "strong_organization_count": 9 }, { "requirement": "identity_delegation", "label": "Identity, access and propagation", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents", "bcg-building-enterprise-ai-agents-in-regulated-industries" ], "source_urls": [ "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale", "https://www.bcg.com/publications/2026/building-enterprise-ai-agents-in-regulated-industries" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-building-the-foundation-for-agentic-ai", "bain-company-the-three-layers-of-an-agentic-ai-platform" ], "source_urls": [ "https://www.bain.com/insights/building-the-foundation-for-agentic-ai-technology-report-2025/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-how-do-you-win-the-game-of-ai-telephone-pwc", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-telephone.html", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Moderate evidence", "source_ids": [ "ey-ey-parthenon-case-study-building-an-enterprise-scale-agentic-ai-os-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_gl/insights/ai/building-an-enterprise-scale-agentic-ai-operating-system", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-invisible-access-visible-risk", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/invisible-access-visible-risk.html", "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-invisible-access-visible-risk", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ], "source_urls": [ "https://www.accenture.com/en/blogs/security/strengthening-ai-agent-security-identity-management", "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Moderate evidence", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Moderate evidence", "source_ids": [ "palantir-action-types-permissions", "palantir-aip-chatbot-studio-overview" ], "source_urls": [ "https://www.palantir.com/docs/foundry/action-types/permissions", "https://www.palantir.com/docs/foundry/chatbot-studio/overview" ], "editorial_note": "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-permissions" ] } }, "interpretation_limits": [ "Technical identity, role/scope checks and identity propagation supported. Formal bounded delegation chains and cryptographic authority attenuation are not specified. the cited source is cultural identity and excluded here.", "IAM/access is better evidenced than portable cross-agent delegation.", "Composite row is overstrong if read as proof of formal bounded delegation, cryptographic attenuation or cross-agent authority chains. Rename to identity/access context or split before publishing." ], "strong_organization_count": 6 }, { "requirement": "pre_effect_authorization", "label": "Permission and approval before action", "assessments": { "McKinsey": { "classification": "Moderate evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ], "source_urls": [ "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish. BCG 08 explicitly discusses blocking calls at the point of execution; it is architectural advice, not a product test.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-what-is-ai-governance", "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank" ], "source_urls": [ "https://www.bain.com/insights/what-is-ai-governance/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/", "https://www.bain.com/insights/what-it-takes-to-build-the-ai-native-modern-bank/" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank" ] }, "PwC": { "classification": "Moderate evidence", "source_ids": [ "pwc-who-s-in-control-managing-the-cyber-risks-of-ai-agents", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://explore.pwc.com/autonomous-ai-in-cyber/ai-control", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Moderate evidence", "source_ids": [ "ey-ey-parthenon-how-ai-use-is-shifting-from-assistive-to-autonomous-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_gl/insights/ai/as-ai-moves-from-advice-to-authority-who-defines-its-limits", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Moderate evidence", "source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value" ], "source_urls": [ "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments", "accenture-systemic-ai-at-the-root-of-manufacturing-performance" ], "source_urls": [ "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments", "https://www.accenture.com/en/insights/digital-engineering-manufacturing/ai-manufacturing" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-how-ai-will-reshape-the-frontlines-of-upstream-oil-and-gas" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2025/oct/how-ai-oil-gas-redefines-competitive-advantage.html" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-action-types-submission-criteria", "palantir-action-types-permissions" ], "source_urls": [ "https://www.palantir.com/docs/foundry/action-types/submission-criteria", "https://www.palantir.com/docs/foundry/action-types/permissions" ], "editorial_note": "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-submission-criteria", "palantir-action-types-permissions" ] } }, "interpretation_limits": [ "Pre-action enforcement and approval conditions supported; durable cryptographic intent commitment before effect is not required by these sources.", "Authorization does not prove completion or eliminate external partial success.", "Retain only for pre-action permissions/approvals. Exact immutable argument commitment, durable one-use intent and confirmed outcome are separate requirements that this grade does not establish." ], "strong_organization_count": 5 }, { "requirement": "human_oversight", "label": "Human oversight and escalation", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-the-agentic-organization-contours-of-the-next-paradigm-for-the-ai-era" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/people%20and%20organizational%20performance/our%20insights/the%20agentic%20organization%20contours%20of%20the%20next%20paradigm%20for%20the%20ai%20era/the-agentic-organization-contours-of-the-next-paradigm-for-the-ai-era.pdf" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-building-enterprise-ai-agents-in-regulated-industries", "bcg-the-agentic-leadership-playbook-a-scaling-strategy-for-ctos-and-cios" ], "source_urls": [ "https://www.bcg.com/publications/2026/building-enterprise-ai-agents-in-regulated-industries", "https://www.bcg.com/publications/2026/agentic-ai-strategy-cio-cto" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-how-to-break-through-pharmas-innovation-bottleneck-with-ai", "bain-company-state-of-the-art-of-agentic-ai-transformation", "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why" ], "source_urls": [ "https://www.bain.com/insights/how-to-break-through-pharmas-innovation-bottleneck-with-ai/", "https://www.bain.com/insights/state-of-the-art-of-agentic-ai-transformation-technology-report-2025/", "https://www.bain.com/insights/your-ai-budget-is-growing-your-returns-arent-heres-why/" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-want-roi-from-ai-go-for-growth-pwc", "pwc-responsible-ai-survey-pwc" ], "source_urls": [ "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-roi.html", "https://www.pwc.com/us/en/tech-effect/ai-analytics/responsible-ai-survey.html" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-responsible-ai-survey-pwc" ] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-the-ai-landscape-in-cybersecurity-ey-us", "ey-ey-parthenon-how-microsoft-uses-genai-for-compliance-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_us/consulting/the-ai-landscape-in-cybersecurity", "https://www.ey.com/en_gl/insights/law/how-microsoft-streamlined-regulatory-compliance-with-gen-ai", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-global-ai-pulse-q2-2026", "kpmg-kpmg-global-ai-in-finance-report", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-pulse.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/kpmg-global-ai-in-finance-report.html", "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-global-ai-pulse-q2-2026", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments", "accenture-systemic-ai-at-the-root-of-manufacturing-performance" ], "source_urls": [ "https://www.accenture.com/en/blogs/security/strengthening-ai-agent-security-identity-management", "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments", "https://www.accenture.com/en/insights/digital-engineering-manufacturing/ai-manufacturing" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-agentic-ai-transforming-compliance-at-financial-institutions", "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/feb/agentic-ai-compliance-reshaping-financial-institutions.html", "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-ai-ethics-and-governance", "palantir-checkpoints-overview" ], "source_urls": [ "https://www.palantir.com/docs/foundry/aip/ethics-governance", "https://www.palantir.com/docs/foundry/checkpoints/overview" ], "editorial_note": "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them. PAL 10 purpose justification must not be described as independent maker-checker approval.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] } }, "interpretation_limits": [ "Earned autonomy, escalation and explicit decision rights recur. Human review has operating cost and cannot be unlimited.", "the cited source is purpose justification; do not claim it is independent maker-checker approval.", "Differentiate approval, monitoring, intervention, justification and independent maker-checker review. A source discussing one does not prove all of them." ], "strong_organization_count": 9 }, { "requirement": "observability_evals", "label": "Observability and evaluation", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents", "bcg-what-happens-when-ai-stops-asking-permission" ], "source_urls": [ "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale", "https://www.bcg.com/publications/2025/what-happens-ai-stops-asking-permission" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-what-is-ai-governance", "bain-company-the-three-layers-of-an-agentic-ai-platform" ], "source_urls": [ "https://www.bain.com/insights/what-is-ai-governance/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-responsible-ai-survey-pwc", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://www.pwc.com/us/en/tech-effect/ai-analytics/responsible-ai-survey.html", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-responsible-ai-survey-pwc", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-ey-expands-nvidia-powered-enterprise-ai-capabilities-with-industry-first-validation-of-nvidia-nemoclaw-for-langchain-ey-global", "ey-ey-parthenon-how-responsible-ai-translates-investment-into-impact-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_gl/newsroom/2026/07/ey-expands-nvidia-powered-enterprise-ai-capabilities-with-industry-first-validation-of-nvidia-nemoclaw-for-langchain", "https://www.ey.com/en_gl/insights/ai/how-can-responsible-ai-bridge-the-gap-between-investment-and-impact", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-global-ai-pulse-q2-2026", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-pulse.html", "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-global-ai-pulse-q2-2026", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-technology-vision-2025-aia-declaration-of-autonomy", "accenture-harnessing-the-power-of-ai-agents", "accenture-state-of-cybersecurity-resilience-2025" ], "source_urls": [ "https://www.accenture.com/content/dam/accenture/final/accenture-com/document-3/Accenture-Tech-Vision-2025.pdf", "https://www.accenture.com/en/insights/ai-data/hive-mind-harnessing-power-ai-agents", "https://www.accenture.com/en/insights/security/state-cybersecurity-2025" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-agentic-ai-transforming-compliance-at-financial-institutions", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/feb/agentic-ai-compliance-reshaping-financial-institutions.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-action-types-action-log", "palantir-aip-evals-overview" ], "source_urls": [ "https://www.palantir.com/docs/foundry/action-types/action-log", "https://www.palantir.com/docs/foundry/aip-evals/overview" ], "editorial_note": "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-action-log", "palantir-aip-evals-overview" ] } }, "interpretation_limits": [ "Continuous production measurement/evaluation is strongly supported. Trace visibility does not expose or prove hidden model cognition.", "Evaluation, event logging, transaction receipts and cryptographic certification differ.", "Composite coverage is not functional proof of both production telemetry and evaluation quality. Neither execution traces nor a source’s reasoning-trace phrase exposes hidden model cognition." ], "strong_organization_count": 9 }, { "requirement": "outcome_reconciliation", "label": "Recovery and external-effect uncertainty", "assessments": { "McKinsey": { "classification": "Weak mention", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-the-telco-reinvention-how-ai-can-fuel-value-creation" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/industries/technology%20media%20and%20telecommunications/telecommunications/our%20insights/the%20telco%20reinvention%20how%20ai%20can%20fuel%20value%20creation/the-telco-reinvention-how-ai-can-fuel-value-creation.pdf" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Weak mention", "source_ids": [ "bcg-how-ai-first-banks-are-rewriting-the-rules-of-retail-banking" ], "source_urls": [ "https://www.bcg.com/publications/2026/how-ai-first-banks-are-rewriting-the-rules-of-retail-banking" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Bain": { "classification": "Weak mention", "source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank", "bain-company-the-architecture-decisions-only-cfos-can-make" ], "source_urls": [ "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/", "https://www.bain.com/insights/what-it-takes-to-build-the-ai-native-modern-bank/", "https://www.bain.com/insights/the-architecture-decisions-only-cfos-can-make/" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank" ] }, "PwC": { "classification": "Weak mention", "source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation. PwC 15 concerns incident eradication and downstream recovery checks, not ambiguous native commit resolution.", "grade_amended_for_public_definition": true, "prior_classification": "Moderate evidence", "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Weak mention", "source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Weak mention", "source_ids": [ "kpmg-kpmg-global-ai-in-finance-report" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/kpmg-global-ai-in-finance-report.html" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Accenture": { "classification": "Weak mention", "source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments" ], "source_urls": [ "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation. Accenture 09 proposes cryptographic payment authority and mentions reconciliation as a service opportunity; it does not specify authoritative resolution of uncertain effects.", "grade_amended_for_public_definition": true, "prior_classification": "Moderate evidence", "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Weak mention", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-agentic-ai-is-reshaping-the-services-industry", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/aug/how-agentic-ai-reshapes-services-firms.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation. Operational exception ownership, review/testing and audit readiness do not specify a native reconciliation algorithm.", "grade_amended_for_public_definition": true, "prior_classification": "Moderate evidence", "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Moderate evidence", "source_ids": [ "palantir-action-types-side-effects-webhooks" ], "source_urls": [ "https://www.palantir.com/docs/foundry/action-types/webhooks" ], "editorial_note": "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation. PAL 05 strongly documents non-atomic external/internal writes; it does not demonstrate a complete authoritative reconciliation capability.", "grade_amended_for_public_definition": true, "prior_classification": "Strong evidence", "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-side-effects-webhooks" ] } }, "interpretation_limits": [ "Data reconciliation, fallback/rollback and recovery appear. No focused evidence establishes authoritative resolution of ambiguous external effects or exactly-once execution as a standalone category.", "Palantir directly documents a partial-success contract; market need is synthesis, not a standalone purchasing survey.", "Narrow authoritative reconciliation is poorly established. Distinguish data reconciliation, incident recovery, deployment rollback, partial-success documentation and native outcome reconciliation." ], "strong_organization_count": 0 }, { "requirement": "security_isolation", "label": "Security, containment and isolation", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-technology-trends-outlook-2025" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/mckinsey%20digital/our%20insights/the%20top%20trends%20in%20tech%202025/mckinsey-technology-trends-outlook-2025.pdf" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-building-enterprise-ai-agents-in-regulated-industries", "bcg-what-happens-when-ai-stops-asking-permission" ], "source_urls": [ "https://www.bcg.com/publications/2026/building-enterprise-ai-agents-in-regulated-industries", "https://www.bcg.com/publications/2025/what-happens-ai-stops-asking-permission" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-what-is-ai-governance", "bain-company-the-three-layers-of-an-agentic-ai-platform" ], "source_urls": [ "https://www.bain.com/insights/what-is-ai-governance/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-who-s-in-control-managing-the-cyber-risks-of-ai-agents", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://explore.pwc.com/autonomous-ai-in-cyber/ai-control", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-ey-expands-nvidia-powered-enterprise-ai-capabilities-with-industry-first-validation-of-nvidia-nemoclaw-for-langchain-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_gl/newsroom/2026/07/ey-expands-nvidia-powered-enterprise-ai-capabilities-with-industry-first-validation-of-nvidia-nemoclaw-for-langchain", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-invisible-access-visible-risk", "kpmg-agentic-ai-advantage-unlocking-next-level-value", "kpmg-cybersecurity-considerations-2025" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/invisible-access-visible-risk.html", "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/cybersecurity-considerations-2025.html" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary. KPMG 05 page 23 explicitly recommends runtime isolation and containment; this supports this narrow concept but remains prescriptive.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-invisible-access-visible-risk", "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-state-of-cybersecurity-resilience-2025" ], "source_urls": [ "https://www.accenture.com/en/blogs/security/strengthening-ai-agent-security-identity-management", "https://www.accenture.com/en/insights/security/state-cybersecurity-2025" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience" ] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-dark-side-of-generative-ai-how-to-navigate-cyber-threats" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2025/may/how-to-navigate-generative-ai-cyberthreat.html" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-aip-security-and-privacy", "palantir-action-types-permissions" ], "source_urls": [ "https://www.palantir.com/docs/foundry/aip/aip-security", "https://www.palantir.com/docs/foundry/action-types/permissions" ], "editorial_note": "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-permissions" ] } }, "interpretation_limits": [ "Least privilege, controlled environments, data protection and containment; isolation details vary.", "Existing zero-trust and platform security investments are encouraged.", "Composite coverage is overstrong if read as proof of runtime/process/network/memory isolation. Rename to security/containment or split. Exact isolation mechanisms and boundaries vary." ], "strong_organization_count": 9 }, { "requirement": "audit_evidence", "label": "Activity records and audit support", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents", "bcg-building-enterprise-ai-agents-in-regulated-industries" ], "source_urls": [ "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale", "https://www.bcg.com/publications/2026/building-enterprise-ai-agents-in-regulated-industries" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-what-is-ai-governance", "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank" ], "source_urls": [ "https://www.bain.com/insights/what-is-ai-governance/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/", "https://www.bain.com/insights/what-it-takes-to-build-the-ai-native-modern-bank/" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-how-do-you-win-the-game-of-ai-telephone-pwc", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ], "source_urls": [ "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-telephone.html", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat" ] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-how-mott-macdonald-accelerated-responsible-ai-ey-global", "ey-ey-parthenon-how-responsible-ai-translates-investment-into-impact-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "source_urls": [ "https://www.ey.com/en_gl/insights/ai/how-mott-macdonald-is-building-confidence-through-responsible-ai", "https://www.ey.com/en_gl/insights/ai/how-can-responsible-ai-bridge-the-gap-between-investment-and-impact", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-kpmg-global-ai-in-finance-report", "kpmg-agentic-ai-advantage-unlocking-next-level-value", "kpmg-ai-in-financial-reporting-and-audit-navigating-the-new-era", "kpmg-guide-ai-and-automation-in-financial-reporting" ], "source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/kpmg-global-ai-in-finance-report.html", "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-in-financial-reporting-and-audit.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-guide-financial-reporting.html" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ], "source_urls": [ "https://www.accenture.com/en/blogs/security/strengthening-ai-agent-security-identity-management", "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-accenture-s-ai-agent-security-enhancing-cyber-resilience", "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-will-reshape-the-frontlines-of-upstream-oil-and-gas", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2025/oct/how-ai-oil-gas-redefines-competitive-advantage.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Strong evidence", "source_ids": [ "palantir-ai-ethics-and-governance", "palantir-action-types-action-log" ], "source_urls": [ "https://www.palantir.com/docs/foundry/aip/ethics-governance", "https://www.palantir.com/docs/foundry/action-types/action-log" ], "editorial_note": "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter. PAL 06 action logs record successful action-type submissions; documentation points to other facilities for failures and edit history.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-action-log" ] } }, "interpretation_limits": [ "Logs, ownership, traceability and reconstructable actions strongly supported. Audit evidence is broader than signed offline execution certificates.", "Scope of records matters; successful-action logs are not all attempted actions or all write paths.", "Logs and audit support are not sufficient appropriate audit evidence by themselves, nor independent signed execution proof. Scope, missing attempts, native records and contradictory evidence matter." ], "strong_organization_count": 9 }, { "requirement": "independent_portable_verification", "label": "Independent portable execution verification", "assessments": { "McKinsey": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage", "mckinsey-company-the-agentic-organization-contours-of-the-next-paradigm-for-the-ai-era", "mckinsey-company-technology-trends-outlook-2025", "mckinsey-company-the-telco-reinvention-how-ai-can-fuel-value-creation" ], "reviewed_tranche_source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/people%20and%20organizational%20performance/our%20insights/the%20agentic%20organization%20contours%20of%20the%20next%20paradigm%20for%20the%20ai%20era/the-agentic-organization-contours-of-the-next-paradigm-for-the-ai-era.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/mckinsey%20digital/our%20insights/the%20top%20trends%20in%20tech%202025/mckinsey-technology-trends-outlook-2025.pdf", "https://www.mckinsey.com/~/media/mckinsey/industries/technology%20media%20and%20telecommunications/telecommunications/our%20insights/the%20telco%20reinvention%20how%20ai%20can%20fuel%20value%20creation/the-telco-reinvention-how-ai-can-fuel-value-creation.pdf" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." }, "BCG": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "bcg-as-ai-investments-surge-ceos-take-the-lead", "bcg-the-widening-ai-value-gap-build-for-the-future-2025", "bcg-from-potential-to-profit-closing-the-ai-impact-gap", "bcg-ai-at-work-momentum-builds-but-gaps-remain", "bcg-ai-at-work-strategy-matters-more-than-tools", "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents", "bcg-building-enterprise-ai-agents-in-regulated-industries", "bcg-why-ai-agents-need-an-identity-not-just-instructions", "bcg-what-happens-when-ai-stops-asking-permission", "bcg-agentic-ai-is-rewriting-the-rules-of-data-risk-management", "bcg-the-agentic-leadership-playbook-a-scaling-strategy-for-ctos-and-cios", "bcg-how-ai-first-banks-are-rewriting-the-rules-of-retail-banking", "bcg-ai-for-healthcare-payers-how-to-become-ai-first", "bcg-ceos-are-starting-to-see-value-from-ai-now-comes-execution", "bcg-the-ai-first-real-estate-company-an-opportunity-for-structural-advantage", "bcg-taking-a-page-from-commodity-traders-a-new-playbook-for-commercial-real-estate-investors", "bcg-rebuilding-asset-management-for-an-ai-first-world", "bcg-distribution-is-the-new-source-of-advantage-in-asset-management", "bcg-an-imperative-for-growth" ], "reviewed_tranche_source_urls": [ "https://www.bcg.com/publications/2026/as-ai-investments-surge-ceos-take-the-lead", "https://media-publications.bcg.com/The-Widening-AI-Value-Gap-Sept-2025.pdf", "https://www.bcg.com/publications/2025/closing-the-ai-impact-gap", "https://www.bcg.com/publications/2025/ai-at-work-momentum-builds-but-gaps-remain", "https://www.bcg.com/publications/2026/ai-at-work-why-strategy-matters-more-than-tools", "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale", "https://www.bcg.com/publications/2026/building-enterprise-ai-agents-in-regulated-industries", "https://www.bcg.com/publications/2026/why-ai-agents-need-an-identity-not-just-instructions", "https://www.bcg.com/publications/2025/what-happens-ai-stops-asking-permission", "https://www.bcg.com/publications/2026/managing-data-risk-in-the-age-of-agentic-ai", "https://www.bcg.com/publications/2026/agentic-ai-strategy-cio-cto", "https://www.bcg.com/publications/2026/how-ai-first-banks-are-rewriting-the-rules-of-retail-banking", "https://www.bcg.com/publications/2026/ai-for-healthcare-payers", "https://www.bcg.com/publications/2026/how-ceos-scale-ai-value", "https://www.bcg.com/publications/2026/the-ai-first-real-estate-company-advantage", "https://www.bcg.com/publications/2026/a-new-playbook-for-commercial-real-estate-investors", "https://www.bcg.com/publications/2026/rebuilding-asset-management-for-an-ai-first-world", "https://www.bcg.com/publications/2026/distribution-is-the-new-source-of-advantage", "https://www.bcg.com/publications/2026/an-imperative-for-growth-and-the-new-economics-of-asset-management" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." }, "Bain": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "bain-company-what-is-ai-governance", "bain-company-better-faster-leaner-reinventing-hr-with-generative-ai", "bain-company-how-to-break-through-pharmas-innovation-bottleneck-with-ai", "bain-company-building-the-foundation-for-agentic-ai", "bain-company-state-of-the-art-of-agentic-ai-transformation", "bain-company-will-agentic-ai-disrupt-saas", "bain-company-from-pilots-to-payoff-generative-ai-in-software-development", "bain-company-the-three-layers-of-an-agentic-ai-platform", "bain-company-why-agentic-ai-demands-a-new-architecture", "bain-company-how-cios-think-about-agentic-ai", "bain-company-your-ai-budget-is-growing-your-returns-aren-t-here-s-why", "bain-company-what-it-takes-to-build-the-ai-native-modern-bank", "bain-company-agentic-ai-commerce-hinges-on-consumer-trust", "bain-company-the-architecture-decisions-only-cfos-can-make", "bain-company-ai-is-transforming-productivity-but-sales-remains-a-new-frontier", "bain-company-ai-wont-save-an-unfocused-game", "bain-company-not-yet-robots-how-to-win-in-medias-flooded-era" ], "reviewed_tranche_source_urls": [ "https://www.bain.com/insights/what-is-ai-governance/", "https://www.bain.com/insights/better-faster-leaner-reinventing-hr-with-generative-ai/", "https://www.bain.com/insights/how-to-break-through-pharmas-innovation-bottleneck-with-ai/", "https://www.bain.com/insights/building-the-foundation-for-agentic-ai-technology-report-2025/", "https://www.bain.com/insights/state-of-the-art-of-agentic-ai-transformation-technology-report-2025/", "https://www.bain.com/insights/will-agentic-ai-disrupt-saas-technology-report-2025/", "https://www.bain.com/insights/from-pilots-to-payoff-generative-ai-in-software-development-technology-report-2025/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/", "https://www.bain.com/insights/why-agentic-ai-demands-a-new-architecture/", "https://www.bain.com/insights/how-cios-think-about-agentic-ai-infographic/", "https://www.bain.com/insights/your-ai-budget-is-growing-your-returns-arent-heres-why/", "https://www.bain.com/insights/what-it-takes-to-build-the-ai-native-modern-bank/", "https://www.bain.com/insights/agentic-ai-commerce-hinges-on-consumer-trust/", "https://www.bain.com/insights/the-architecture-decisions-only-cfos-can-make/", "https://www.bain.com/insights/ai-transforming-productivity-sales-remains-new-frontier-technology-report-2025/", "https://www.bain.com/insights/ai-wont-save-an-unfocused-game-gaming-report-2026/", "https://www.bain.com/insights/not-yet-robots-how-to-win-in-medias-flooded-era/" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." }, "PwC": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "pwc-decoding-roi-from-ai-pwc", "pwc-the-rise-of-autonomous-ai-in-cybersecurity-pwc", "pwc-2026-ai-business-predictions-pwc", "pwc-how-do-you-win-the-game-of-ai-telephone-pwc", "pwc-how-ai-is-transforming-research-and-development-pwc", "pwc-pwc-launches-ai-agent-operating-system-for-enterprises-pwc", "pwc-want-roi-from-ai-go-for-growth-pwc", "pwc-responsible-ai-survey-pwc", "pwc-maximising-the-impact-of-agentic-ai-pwc", "pwc-ai-in-operations-pwc", "pwc-nine-ai-fuelled-business-models-that-leaders-cant-ignore-pwc", "pwc-who-s-in-control-managing-the-cyber-risks-of-ai-agents", "pwc-agents-of-change-when-ai-agents-lose-their-way-how-to-manage-the-insider-threat", "pwc-rethinking-learning-in-the-age-of-ai-pwc", "pwc-ai-agent-survey-pwc", "pwc-pwcs-29th-global-ceo-survey-leading-through-uncertainty-in-the-age-of-ai", "pwc-us-ai-jobs-barometer-pwc" ], "reviewed_tranche_source_urls": [ "https://www.pwc.com/gx/en/1/issues/tech-data-ai/decoding-ai-roi.html", "https://www.pwc.com/gx/en/1/issues/cybersecurity/ai-cybersecurity.html", "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-predictions.html", "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-telephone.html", "https://www.pwc.com/gx/en/1/services/ai/ai-research-development.html", "https://www.pwc.com/us/en/about-us/newsroom/press-releases/pwc-launches-ai-agent-operating-system-enterprises.html", "https://www.pwc.com/gx/en/1/issues/tech-data-ai/ai-roi.html", "https://www.pwc.com/us/en/tech-effect/ai-analytics/responsible-ai-survey.html", "https://www.pwc.com/gx/en/1/issues/c-suite-insights/the-leadership-agenda/ai-agents-survey.html", "https://www.pwc.com/gx/en/1/services/consulting/technology/ai-operations.html", "https://www.pwc.com/gx/en/1/issues/reinvention/ai-nine-business-models.html", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-control", "https://explore.pwc.com/autonomous-ai-in-cyber/ai-agents-part-2", "https://www.pwc.com/gx/en/1/issues/tech-data-ai/next-tech-agenda/ai-learning.html", "https://www.pwc.com/us/en/tech-effect/ai-analytics/ai-agent-survey.html", "https://www.pwc.com/gx/en/1/issues/c-suite-insights/ceo-survey.html", "https://www.pwc.com/us/en/tech-effect/ai-analytics/ai-jobs-barometer.html" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." }, "EY / EY-Parthenon": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "ey-ey-parthenon-how-ai-use-is-shifting-from-assistive-to-autonomous-ey-global", "ey-ey-parthenon-ceo-priorities-2026-growth-resilience-and-ai-roi-ey-global", "ey-ey-parthenon-case-study-building-an-enterprise-scale-agentic-ai-os-ey-global", "ey-ey-parthenon-responsible-ai-principles-for-competitive-advantage-ey-global", "ey-ey-parthenon-ey-expands-nvidia-powered-enterprise-ai-capabilities-with-industry-first-validation-of-nvidia-nemoclaw-for-langchain-ey-global", "ey-ey-parthenon-the-ai-landscape-in-cybersecurity-ey-us", "ey-ey-parthenon-can-ai-advance-toward-value-if-workforce-tensions-linger-ey-global", "ey-ey-parthenon-agentic-ai-next-frontier-of-human-and-ai-orchestration-ey-global", "ey-ey-parthenon-how-ai-innovation-powers-microsofts-finance-journey-ey-global", "ey-ey-parthenon-how-ai-drove-data-optimization-for-oil-and-gas-capital-projects-ey-global", "ey-ey-parthenon-how-mott-macdonald-accelerated-responsible-ai-ey-global", "ey-ey-parthenon-how-microsoft-uses-genai-for-compliance-ey-global", "ey-ey-parthenon-case-study-how-ai-automated-insurance-claims-ey-global", "ey-ey-parthenon-how-ai-helped-a-health-care-company-transform-delivery-ey-global", "ey-ey-parthenon-how-novartis-optimized-its-supply-chain-ey-global", "ey-ey-parthenon-how-responsible-ai-translates-investment-into-impact-ey-global", "ey-ey-parthenon-agentic-ai-set-to-boost-global-infrastructure-productivity-new-ey-report-ey-global", "ey-ey-parthenon-agentic-ai-governance-and-real-time-trust-ey-us" ], "reviewed_tranche_source_urls": [ "https://www.ey.com/en_gl/insights/ai/as-ai-moves-from-advice-to-authority-who-defines-its-limits", "https://www.ey.com/en_gl/ceo/ceo-outlook-global-report", "https://www.ey.com/en_gl/insights/ai/building-an-enterprise-scale-agentic-ai-operating-system", "https://www.ey.com/en_gl/insights/ai/principles-for-ethical-and-responsible-ai", "https://www.ey.com/en_gl/newsroom/2026/07/ey-expands-nvidia-powered-enterprise-ai-capabilities-with-industry-first-validation-of-nvidia-nemoclaw-for-langchain", "https://www.ey.com/en_us/consulting/the-ai-landscape-in-cybersecurity", "https://www.ey.com/en_gl/insights/workforce/work-reimagined-survey", "https://www.ey.com/en_gl/services/ai/agentic-ai", "https://www.ey.com/en_gl/insights/consulting/how-ai-innovation-powers-microsoft-s-finance-journey", "https://www.ey.com/en_gl/insights/energy-resources/how-ai-drove-data-optimization-for-oil-and-gas-capital-projects", "https://www.ey.com/en_gl/insights/ai/how-mott-macdonald-is-building-confidence-through-responsible-ai", "https://www.ey.com/en_gl/insights/law/how-microsoft-streamlined-regulatory-compliance-with-gen-ai", "https://www.ey.com/en_gl/insights/financial-services/emeia/how-a-nordic-insurance-company-automated-claims-processing", "https://www.ey.com/en_gl/services/technology/how-one-health-care-testing-company-gets-results-with-ai", "https://www.ey.com/en_gl/alliances/novartis-optimizes-supply-chain-with-real-time-risk-insight", "https://www.ey.com/en_gl/insights/ai/how-can-responsible-ai-bridge-the-gap-between-investment-and-impact", "https://www.ey.com/en_gl/newsroom/2026/05/agentic-ai-set-to-boost-global-infrastructure-productivity-new-ey-report", "https://www.ey.com/en_us/insights/ai/agentic-ai-governance-and-real-time-trust" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." }, "KPMG": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "kpmg-global-ai-pulse-q2-2026", "kpmg-kpmg-global-ai-in-finance-report", "kpmg-global-tech-report-2026", "kpmg-invisible-access-visible-risk", "kpmg-agentic-ai-advantage-unlocking-next-level-value", "kpmg-ai-s-dual-promise-enabling-climate-outcomes-and-powering-the-energy-transition", "kpmg-kpmg-and-microsoft-scale-trusted-enterprise-ai-agents-globally-through-deployment-of-agent-365-and-copilot", "kpmg-kpmg-managed-services-outlook-survey-2026", "kpmg-trust-attitudes-and-use-of-artificial-intelligence-a-global-study-2025", "kpmg-cybersecurity-considerations-2025", "kpmg-ai-in-legal", "kpmg-global-ai-pulse-q1-2026", "kpmg-ai-in-financial-reporting-and-audit-navigating-the-new-era", "kpmg-guide-ai-and-automation-in-financial-reporting", "kpmg-strengthening-trust-and-regulatory-compliance-with-ai" ], "reviewed_tranche_source_urls": [ "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-pulse.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/kpmg-global-ai-in-finance-report.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/global-tech-report.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/invisible-access-visible-risk.html", "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html", "https://kpmg.com/xx/en/our-insights/esg/ai-enabling-climate-outcomes-and-powering-the-energy-transition.html", "https://kpmg.com/xx/en/media/press-releases/2026/06/kpmg-and-microsoft-scale-trusted-enterprise-ai-agents-globally.html", "https://kpmg.com/xx/en/our-insights/transformation/managed-services-outlook.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/trust-attitudes-and-use-of-ai.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/cybersecurity-considerations-2025.html", "https://kpmg.com/xx/en/what-we-do/services/law/ai-in-legal.html", "https://assets.kpmg.com/content/dam/kpmgsites/xx/pdf/2026/04/global-ai-pulse.pdf", "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-in-financial-reporting-and-audit.html", "https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-guide-financial-reporting.html", "https://kpmg.com/xx/en/what-we-do/services/kpmg-client-stories/regulatory-change/strengthening-trust-and-regulatory-compliance-with-ai.html" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." }, "Accenture": { "classification": "No meaningful evidence", "source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments" ], "source_urls": [ "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments" ], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions. No selected passage establishes independent/offline execution-proof portability. The original W marks adjacent ideas, not the named function.", "grade_amended_for_public_definition": true, "prior_classification": "Weak mention", "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "No meaningful evidence", "source_ids": [ "oliver-wyman-why-trust-gives-banks-an-edge-in-the-ai-economy", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/jan/why-trust-gives-banks-edge-in-ai-economy.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions. No selected passage establishes independent/offline execution-proof portability. The original W marks adjacent ideas, not the named function.", "grade_amended_for_public_definition": true, "prior_classification": "Weak mention", "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "No meaningful evidence", "source_ids": [], "source_urls": [], "editorial_note": "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [], "reviewed_tranche_source_ids": [ "palantir-ai-ethics-and-governance", "palantir-aip-security-and-privacy", "palantir-action-types-submission-criteria", "palantir-action-types-permissions", "palantir-action-types-side-effects-webhooks", "palantir-action-types-action-log", "palantir-aip-evals-overview", "palantir-overview-ontology", "palantir-checkpoints-overview", "palantir-cleveland-clinic-palantir", "palantir-sompo-and-palantir", "palantir-impact-airbus-and-skywise", "palantir-aip-chatbot-studio-overview" ], "reviewed_tranche_source_urls": [ "https://www.palantir.com/docs/foundry/aip/ethics-governance", "https://www.palantir.com/docs/foundry/aip/aip-security", "https://www.palantir.com/docs/foundry/action-types/submission-criteria", "https://www.palantir.com/docs/foundry/action-types/permissions", "https://www.palantir.com/docs/foundry/action-types/webhooks", "https://www.palantir.com/docs/foundry/action-types/action-log", "https://www.palantir.com/docs/foundry/aip-evals/overview", "https://www.palantir.com/docs/foundry/ontology/overview", "https://www.palantir.com/docs/foundry/checkpoints/overview", "https://www.palantir.com/impact/cleveland-clinic/", "https://www.palantir.com/impact/sompo/", "https://www.palantir.com/impact/airbus/", "https://www.palantir.com/docs/foundry/chatbot-studio/overview" ], "absence_basis": "No qualifying positive evidence was established in this reviewed tranche. These links identify the assessed collection; they are not positive proof of this function or evidence of organization-wide absence." } }, "interpretation_limits": [ "No reviewed source identifies independent/offline signed action-proof portability as an urgent separate buyer requirement. This is absence of evidence in this tranche, not proof of no market.", "Cryptographic intent, verification preference and tamper-evident trails do not establish offline execution-certificate budgets.", "No meaningful evidence means not established in the reviewed tranche, not organization-wide absence, opposition, or an empty market. Cryptographic intent and model portability are different functions." ], "strong_organization_count": 0 }, { "requirement": "interoperability", "label": "Interfaces and interoperability", "assessments": { "McKinsey": { "classification": "Strong evidence", "source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale", "mckinsey-company-seizing-the-agentic-ai-advantage" ], "source_urls": [ "https://www.mckinsey.com/~/media/mckinsey/business%20functions/technology/our%20insights/building%20the%20foundations%20for%20agentic%20ai%20at%20scale/building-the-foundations-for-agentic-ai-at-scale_vf.pdf", "https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/seizing%20the%20agentic%20ai%20advantage/seizing-the-agentic-ai-advantage.pdf" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only. McKinsey coverage is only five counted publications; do not extrapolate this row to all of the organization’s work.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "mckinsey-company-building-the-foundations-for-agentic-ai-at-scale" ] }, "BCG": { "classification": "Strong evidence", "source_ids": [ "bcg-the-widening-ai-value-gap-build-for-the-future-2025", "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents", "bcg-building-enterprise-ai-agents-in-regulated-industries" ], "source_urls": [ "https://media-publications.bcg.com/The-Widening-AI-Value-Gap-Sept-2025.pdf", "https://www.bcg.com/publications/2026/how-cios-govern-ai-agents-at-scale", "https://www.bcg.com/publications/2026/building-enterprise-ai-agents-in-regulated-industries" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bcg-enterprise-ai-control-plane-the-cios-guide-to-governing-and-accelerating-ai-agents" ] }, "Bain": { "classification": "Strong evidence", "source_ids": [ "bain-company-building-the-foundation-for-agentic-ai", "bain-company-will-agentic-ai-disrupt-saas", "bain-company-the-three-layers-of-an-agentic-ai-platform" ], "source_urls": [ "https://www.bain.com/insights/building-the-foundation-for-agentic-ai-technology-report-2025/", "https://www.bain.com/insights/will-agentic-ai-disrupt-saas-technology-report-2025/", "https://www.bain.com/insights/the-three-layers-of-an-agentic-ai-platform/" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "bain-company-will-agentic-ai-disrupt-saas", "bain-company-the-three-layers-of-an-agentic-ai-platform" ] }, "PwC": { "classification": "Strong evidence", "source_ids": [ "pwc-pwc-launches-ai-agent-operating-system-for-enterprises-pwc" ], "source_urls": [ "https://www.pwc.com/us/en/about-us/newsroom/press-releases/pwc-launches-ai-agent-operating-system-enterprises.html" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "EY / EY-Parthenon": { "classification": "Strong evidence", "source_ids": [ "ey-ey-parthenon-case-study-building-an-enterprise-scale-agentic-ai-os-ey-global" ], "source_urls": [ "https://www.ey.com/en_gl/insights/ai/building-an-enterprise-scale-agentic-ai-operating-system" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Source inventory and claim record reviewed; not all original passages revisited for this report", "rechecked_source_ids": [] }, "KPMG": { "classification": "Strong evidence", "source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value", "kpmg-kpmg-and-microsoft-scale-trusted-enterprise-ai-agents-globally-through-deployment-of-agent-365-and-copilot", "kpmg-kpmg-managed-services-outlook-survey-2026" ], "source_urls": [ "https://kpmg.com/xx/en/what-we-do/services/ai/agentic-ai-advantage.html", "https://kpmg.com/xx/en/media/press-releases/2026/06/kpmg-and-microsoft-scale-trusted-enterprise-ai-agents-globally.html", "https://kpmg.com/xx/en/our-insights/transformation/managed-services-outlook.html" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "kpmg-agentic-ai-advantage-unlocking-next-level-value" ] }, "Accenture": { "classification": "Strong evidence", "source_ids": [ "accenture-building-a-reinvention-ready-digital-core", "accenture-agentic-commerce-and-the-future-of-payments", "accenture-the-new-rules-of-platform-strategy-in-the-age-of-agentic-ai" ], "source_urls": [ "https://www.accenture.com/en/insights/technology/building-reinvention-ready-digital-core", "https://www.accenture.com/en/blogs/banking/agentic-commerce-payments", "https://www.accenture.com/en/insights/strategy/new-rules-platform-strategy-agentic-ai" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "accenture-agentic-commerce-and-the-future-of-payments" ] }, "Oliver Wyman": { "classification": "Strong evidence", "source_ids": [ "oliver-wyman-how-to-use-agentic-ai-to-boost-procurement-efficiency", "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ], "source_urls": [ "https://www.oliverwyman.com/our-expertise/insights/2026/mar/agentic-ai-transform-procurement-operating-model.html", "https://www.oliverwyman.com/our-expertise/insights/2026/may/ai-for-legacy-it.html", "https://www.oliverwyman.com/our-expertise/insights/2026/sep/agentic-ai-enterprise-operating-model.html" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "oliver-wyman-how-ai-for-enterprise-legacy-it-runs-operations-at-scale", "oliver-wyman-why-it-operating-models-need-to-change-before-agentic-ai" ] }, "Palantir": { "classification": "Moderate evidence", "source_ids": [ "palantir-action-types-side-effects-webhooks", "palantir-overview-ontology", "palantir-aip-chatbot-studio-overview" ], "source_urls": [ "https://www.palantir.com/docs/foundry/action-types/webhooks", "https://www.palantir.com/docs/foundry/ontology/overview", "https://www.palantir.com/docs/foundry/chatbot-studio/overview" ], "editorial_note": "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only.", "grade_amended_for_public_definition": false, "prior_classification": null, "evidence_scope": "Selected passages revisited", "rechecked_source_ids": [ "palantir-action-types-side-effects-webhooks", "palantir-overview-ontology" ] } }, "interpretation_limits": [ "Hybrid estates and stable interfaces recur; Bain explicitly warns MCP/A2A syntax does not supply shared business semantics, and vendor incentives resist open universality.", "Native platforms expose APIs while retaining context and control ownership.", "Stable interfaces and integration do not prove shared business semantics, portable policy, outcome receipts, or vendor-neutral control. Retain as editorial theme coverage only." ], "strong_organization_count": 8 } ] }