THE PROBLEM WITH THE MARKET
In our view, AI advisors rarely compete on methodology. They compete on relationship. Most institutions engage the firm they already use for audit, strategy, or technology — and tend to adopt the framing that firm brings.
Framing is not neutral.
A consulting firm aligned with a hyperscaler may naturally be more familiar with, and experienced in, that infrastructure ecosystem. A firm whose own operations run on a partner model vendor may be more inclined to build around technologies it already understands and uses. A software company that sells AI evaluates AI through the lens of its own platform and capabilities.
The point is not ethics. It is economic structure, operating models, and commercial alignment — and those structures are what we analyze below.
THE FIVE ARCHETYPES
01 · THE FRAMEWORK CONSULTANCY
What they sell. Board roadmaps, governance frameworks, responsible-AI principles, operating models, and policy guidance. The deliverables are policies, committees, accountability structures, and decision frameworks.
Economic structure. The model is designed to help institutions define governance expectations, risk-management approaches, and organizational responsibilities. Its primary focus is establishing how AI should be governed rather than how technical controls are implemented and enforced within the underlying infrastructure.
Where the model reaches its limit. Governance guidance alone does not create technical control. Institutions must still translate policies, standards, and oversight objectives into enforceable architectural and operational controls.
02 · THE HYPERSCALER-ALIGNED STRATEGIST
What they sell. Enterprise strategy paired with delivery built around a cloud platform ecosystem — often supported by joint teams, co-funded initiatives, and aligned go-to-market programs.
Economic structure. The firm's experience, delivery capabilities, and commercial relationships are frequently concentrated within a particular infrastructure ecosystem. As a result, strategic recommendations may naturally reflect the environments in which the firm has the deepest expertise and implementation experience.
Where the model reaches its limit. The model is optimized to help institutions succeed within a particular ecosystem. Evaluating Build, Rent, Compose, or Rent + Control from a fully ecosystem-independent perspective may require a broader architectural lens.
03 · THE MODEL-VENDOR-FUSED ADVISOR
What they sell. AI transformation programs built around a foundation-model ecosystem, with client delivery, internal tooling, and operational workflows often centered on the same technology stack.
Economic structure. The firm's internal operations, delivery methodologies, and accumulated expertise are frequently tied to a particular model ecosystem. As a result, recommendations may naturally build upon the platforms, tools, and operating environments with which the firm has the greatest experience.
Where the model reaches its limit. Strategic guidance is often developed within the context of a specific model ecosystem, which can make broader architectural alternatives more difficult to evaluate on equal footing.
04 · THE AUDIT-ANCHORED ADVISORY
What they sell. AI inventories, risk classification, testing, monitoring, governance reviews, and accountability frameworks — traditional risk-management and assurance disciplines extended to AI.
Economic structure. The model is designed to identify, assess, monitor, and report on AI-related risks. These activities provide visibility into AI use and governance posture, but they are distinct from designing and operating the technical controls embedded within the underlying infrastructure.
Where the model reaches its limit. Risk oversight and assurance help institutions understand AI risk. Implementing and enforcing infrastructure-level control requires a complementary architectural and operational discipline.
05 · THE control SOFTWARE VENDOR
What they sell. Software that monitors, logs, analyzes, and reports on AI activity, typically operating within a broader AI, data, or infrastructure platform.
Economic structure. The model is designed to provide visibility, governance, and operational insight within the platform ecosystem in which it operates. Monitoring, reporting, and control capabilities are often closely integrated with the underlying technology stack.
Where the model reaches its limit. Platform-native oversight provides deep visibility into a specific environment. Institutions seeking a broader view across multiple platforms, providers, or operating environments may require an additional layer of architectural assessment and governance.
WHAT THIS MEANS FOR THE INSTITUTION
The archetypes are not flawed. They are designed to solve different problems.
A framework consultancy can help a board establish governance structures and decision frameworks. A hyperscaler-aligned strategist can help institutions deploy and scale within a particular cloud ecosystem. A model-focused advisor can accelerate adoption within a specific AI environment. An audit and risk advisory can help inventory, assess, and monitor AI-related risks. A control software provider can deliver visibility and governance capabilities within its platform.
In our analysis, each model brings valuable capabilities to the institution. The Build, Rent, Compose, and Rent + Control decisions, however, often require a broader architectural assessment that spans multiple ecosystems, providers, operating models, and control approaches.
Institutional AI is typically engaged when an institution seeks an independent evaluation of those architectural choices and their implications for control, governance, resilience, and long-term ownership. The archetypes may continue to play an important role afterward — helping implement, operate, monitor, and scale the direction the institution has selected under a control architecture it owns.
HOW TO TEST THE CATEGORY
A simple diligence test.
These are reasonable questions to ask any AI advisor — including us. The answers tend to reveal how a firm's economics relate to the advice it gives.
- Does your recommendation depend on a specific cloud platform, model vendor, or software product?
- How is your firm's revenue related to any single infrastructure or model partner?
- If we concluded that Build was the right answer for our most sensitive workloads, how would that affect your engagement economics?
- Can you produce a control architecture that remains the institution's, independent of any continuing relationship with you?
There are good answers to each. What matters is that the institution asks.