A consulting firm can build the best advisory agent in its market and still fail to own the business around it. The agent may run inside the client's environment. The client may control the data and usage history. A model provider may control discovery. An enterprise software company may own the workflow. The consulting firm may supply the method and receive none of the information that would make the next deployment better.
The commercial value of owning the advisory product therefore depends on which surrounding rights the firm can retain. That distinction was missing from my 2022 proposal for a consulting app store. I assumed the firm would own the environment through which clients accessed its research, tools, agents and experts.
That remains one possible outcome, but the client and technology providers are also credible owners of that environment. The firm needs a strategy that accounts for all three.
Four parties are moving toward the same workflow
Consulting firms are building reusable AI assets. McKinsey describes Lilli and client versions of its underlying architecture. Deloitte describes Zora AI as a multi-agent offering for business processes. Some firms are helping clients own the environment instead. IBM says Enterprise Advantage helps organisations build and operate their own tailored internal AI platforms while retaining existing cloud providers, models and infrastructure.
Model providers are moving closer to implementation. OpenAI launched a deployment company in May 2026. Anthropic's partner network supports firms helping enterprises move Claude into production. Enterprise software companies already own many of the systems where work occurs. Microsoft Copilot Studio and Salesforce Agentforce put agent creation and operation inside their existing platforms.
Four parties are converging on the same enterprise workflow: the consulting firm, the model provider, the software platform and the client. The competition extends beyond application quality to control of the environment in which the application is discovered, connected to client context and used repeatedly. Those decisions affect who captures the economics and learns from deployment.
The Four Rights
Discussing platform ownership as a single asset obscures the rights that determine the consulting firm's position. A firm can own the intellectual property while another party controls distribution, access to client context or the learning generated through use.
Recent work on vertical AI calls one version of this "going headless": the domain capability survives while the interface moves to an agent or orchestrator. The market is breaking ownership into The Four Rights.
1. The right to distribute
Who controls where the client discovers, invokes and combines the capability? A consulting-owned platform keeps the firm between the product and the client. A module running inside a client's environment or an enterprise software marketplace does not. The firm may own the code while another party controls ranking, bundling, substitution and the customer relationship.
A strong product can therefore lose commercial relevance if another party controls whether clients discover or invoke it.
2. The right to use client context
Who can reach the data, permissions, workflow state and decision history required to produce a useful answer? Context tends to accumulate inside the systems where work already happens. In Accenture's 2025 survey of 1,031 executives, 66 per cent said they relied primarily on platform-native AI; 32 per cent were building platform-agnostic agents across systems.
A consulting portal can contain exceptional knowledge and still sit outside the operational context that makes it valuable every day.
3. The right to improve
Who receives the corrections, edge cases, outcome evidence and usage patterns generated by each deployment? Improvement rights deserve particular attention because they determine whether repeated deployment strengthens the firm's own capability. A portable product can create value for the client while teaching the platform owner more than it teaches the consulting firm. Over time, the firm supplies the method while somebody else accumulates the learning.
The right can be preserved without owning the interface. Contracts can provide for client-authorised feedback, evaluation results or privacy-preserving telemetry. But it has to be designed. It does not arrive merely because the firm owns the original intellectual property.
4. The right to endorse
Who is prepared to say that the output is good enough for the decision being made? This right sits outside software ownership. Deloitte's Zora AI materials separate the technology from client-specific control design, named owners, human oversight and ongoing monitoring. The agent can generate the output. The surrounding operating model decides who stands behind it.
For low-consequence work, endorsement may add little. For a board, regulatory or investor decision, it may be the most valuable right in the bundle.
Three futures
The four rights produce three broad market structures.
The consulting firm owns the environment. It controls distribution, maintains the direct client relationship and may operate the improvement loop. This is the closest version of the consulting app store.
The technology provider owns the environment. The model provider or enterprise software company controls the platform. Consulting firms supply implementation, industry knowledge, workflow design and accountable review inside it.
The client owns the environment. The enterprise controls identity, permissions, data, models and workflow history. It buys diagnostics, agents, benchmarks, integration and expert review from several providers.
A single firm may work across all three. The question is which rights it retains in each.
How a platform strategy disappoints
A platform can work technically and still fail strategically.
It distributes generic content. Frameworks, templates and synthesised research create convenience but little advantage when a general-purpose model can approximate them.
It sits outside the workflow. The client visits for occasional analysis but does repeated work somewhere else.
It loses the learning loop. Corrections and outcomes stay with the client or platform owner, so the product does not compound with use.
It leaves endorsement undefined. The client may not be able to tell what the firm reviewed, what it stands behind or when an expert will intervene.
It avoids the cannibalisation decision. A reusable product reduces work that once supported a project team, but the business case assumes the new revenue is additive.
Each failure can occur even when the technology performs well. The platform strategy must therefore specify how the firm retains the rights needed for its chosen business model.
The investment test
Before funding a consulting platform, ask:
Which rights do we need? Distribution, context, improvement and endorsement create different businesses.
Why will the client grant them? Firm ambition is not evidence of client preference.
What do we learn from use? If the answer is nothing, the product may improve the client's capability faster than the firm's.
Can the capability survive outside our interface? Portability may sacrifice distribution control but preserve relevance when clients own the environment.
What are we prepared to stand behind? The product should make the answer explicit before the first consequential output.
What revenue are we willing to shrink? If the platform works, some traditional work should become smaller. That belongs in the strategy.
As consulting expertise becomes modular, it becomes easier to deploy inside an environment owned by the client or a technology provider. That can expand the firm's reach, but it also separates ownership of the method from control of the interface, context and learning created through use.
Before funding the platform, leaders should identify which rights support the proposed economics and how each will be preserved through product design, integration and contracting. A firm-owned consulting app store may emerge, but the more immediate requirement is a business model that remains valuable across the environments clients actually choose.