01 Compare the argument, then make a choice
This is a qualitative comparison of selected accessible publications, not a ranking of firm capabilities or an exhaustive census. It extends the standalone 30-firm CFO comparison with enterprise organization, executive partnering and capability design.
Different evidence deserves different weight. Advisory frameworks suggest designs; surveys describe respondents; cases may report ambitions rather than achieved results; investor theses identify possible profit pools; standards and governance frameworks guide practice. A public abstract supports only its visible claims.
02 Professional services: What each selected view emphasizes
| Firm | Distinct emphasis in reviewed material | Leadership and capability implication | Koko's added question | Evidence |
|---|---|---|---|---|
| McKinsey | Agentic organization spanning business models, outcome teams, governance, workforce and technology. | Treat operating-model redesign as an executive agenda. | Which capability demonstrably improves enterprise value after coordination and control costs? | McKinsey ↗, McKinsey ↗ |
| BCG | Explicit operating logic; customer-journey teams; common platforms and controls. | Join distributed outcome ownership with shared foundations. | Can the enterprise test a decision rule, reconstruct an outcome and change a supplier? | BCG ↗, BCG ↗, BCG ↗ |
| Deloitte | Continuous orchestration of skills and capacity; human/AI work design; procurement adaptation. | Link workforce choices to operating outcomes as conditions change. | Is the learning path producing the judgment the new model requires? | Deloitte ↗, Deloitte ↗, Deloitte ↗ |
| PwC | Integrated customer journeys and broader economics for agentic global business services. | Connect commercial decisions and replace labor-only comparisons with total capability economics. | Does apparent efficiency survive revenue leakage, transition costs and service accountability? | PwC ↗, PwC ↗ |
| EY | CIO-led orchestration and knowledge infrastructure; capability-center redesign. | Connect the agent layer to cross-functional delivery. | Which interfaces and management activities can actually change without weakening judgment or recovery? | EY ↗, EY ↗ |
| KPMG | Technology convergence and an evolving CISO resilience mandate. | Connect frontier scanning with dependable operating and security foundations. | Which development changes a real capability decision now, and which remains an option? | KPMG ↗, KPMG ↗, KPMG ↗ |
| Accenture | Shared business/technology capability, modular architecture and operating-model reinvention. | Fund usable common foundations alongside outcome delivery. | Who consumes each foundation, and does reuse improve the economics? | Accenture ↗, Accenture ↗ |
| Capgemini | Trust, human/AI collaboration and deliberate authorization of operational agents. | Make scope, ownership and controls part of deployment. | Can authority be enforced during a difficult exception, beyond a documented policy? | Capgemini ↗, World Economic Forum ↗, Capgemini ↗ |
| Oliver Wyman | Gaps among ambition, foundation funding, governance and talent. | Rebuild the conditions for productive agent use. | What should stop receiving marginal funding until the bottleneck is repaired? | Oliver Wyman ↗ |
| Kearney | AI as a business capability, supported by enterprise lifecycle governance and value monitoring. | Connect executive intent with operational ownership. | What is the smallest complete capability that can prove value and control? | Kearney ↗, Kearney ↗ |
| Bain | Historical operating-model principles connecting strategy, decisions and accountabilities. | Start with strategic choices before reorganizing. | Would this design be useful even without the AI label? | Bain ↗ |
Bain’s entry is explicitly a historical foundation, not a substitute for an unseen current agentic report. Selected PDF excerpts for Kearney, Accenture and EY carry narrower access than a full-report review. Other finance-focused firms remain represented in the connected CFO comparison; their enterprise-specific positions are not inferred from finance publications.
03 Analysts, investors and leadership advisers
| Source | What it adds | Boundary and Koko implication |
|---|---|---|
| Forrester | CIO outcome governance and a CDO mandate for trusted context. Forrester ↗, Forrester ↗ | The CDO report is abstract-only. Specify the partnership without pretending one executive owns all AI decisions. |
| Gartner | CIO participation in enterprise work-resource design and business-led agent governance. Gartner ↗, Gartner ↗ | Indexed excerpts/abstracts with restricted direct access. Treat forecasts as forecasts and avoid claiming unseen findings. |
| IDC | Executive pressure to connect AI to outcomes, efficiency and risk. IDC ↗ | Public hub summaries span studies. Do not pool their samples or convert priorities into a causal model. |
| Andreessen Horowitz | Maintained context and the continuing power of incumbent systems. Andreessen Horowitz ↗, Andreessen Horowitz ↗ | Investor perspective. Test both context investment and switching economics. |
| Foundation Capital | Decision records and software designed for agent operators. Foundation Capital ↗, Foundation Capital ↗ | Investor thesis. A record of a decision is not evidence that the decision was correct or authorized. |
| Sequoia | Services as an alternative market for AI-delivered work. Sequoia Capital ↗ | Investment thesis. Purchasing an outcome shifts responsibilities that contracts and operations must resolve. |
| Bessemer | Evaluation of AI-native services and ownership of outcomes. Bessemer Venture Partners ↗ | Investor framework. Autonomous capability does not guarantee an attractive service business. |
| Russell Reynolds / Spencer Stuart | Role fit, executive alignment and distinct technology mandates. Russell Reynolds Associates ↗, Spencer Stuart ↗ | Historical leadership guidance. Design the mandate and relationships before deciding the title. |
04 Where the views agree—and where agreement stops
There is directional convergence around redesigning work, shared data and technology foundations, human development and explicit governance. That is useful support for experimentation. It is not empirical proof of a particular organization chart.
The unresolved questions are more consequential: how much autonomy is dependable; whether coordination becomes cheaper after control costs; how tacit expertise is preserved; where platform bargaining power settles; and which capabilities should be internal or external.
Koko should be willing to reject fashionable conclusions. A smaller workforce is not the same as a stronger enterprise. More agents are not the same as more capability. A documented governance framework is not the same as an enforced boundary. A rapid demonstration is not a repeatable operating result.
05 Evidence that challenges the narrative
The P&G experiment supports some cross-disciplinary benefits of AI in bounded innovation work. METR’s historical experiment, later methodological update and newer survey show why productivity claims require attention to task, population, technology date and measurement. These studies examine different settings and should not be averaged. NBER ↗ METR ↗ METR ↗ METR ↗
Microsoft’s Work Chart narrative references the P&G research; that reference is not a second independent experiment. Deloitte’s human-capital overview and orchestration chapter are one research family. The WEF/Capgemini playbook is joint work. The source count measures coverage, not independent votes for Koko’s thesis.
06 What should change the next edition?
Prioritize evidence of sustained outcomes in real operating environments: controls under stress, capability reuse across domains, preserved or improved professional quality, measured apprenticeship outcomes, and credible supplier switching.
Seek more named longitudinal cases and sector-specific counterexamples. Expand enterprise-specific coverage of firms currently represented only in the CFO comparison. Update a thesis when the evidence changes its rationale, confidence, signpost or associated investment. Retain the old position and explain the revision.