Things are changing fast.

OpenAI launched the OpenAI Deployment Company, a $4 billion consulting venture backed by TPG, Bain Capital, Brookfield, McKinsey, and 15 other investment and consulting firms. Anthropic announced a parallel services operation backed by Blackstone, Goldman Sachs, Hellman and Friedman, and Sequoia. Google Cloud committed $750 million to fund consulting firms, including McKinsey, Accenture, and Deloitte, to build and sell agentic AI projects using Gemini, with Google engineers embedded in the partnerships.

Combined, the two leading AI labs committed over $11 billion to consulting operations within a single day.

The AI labs just became consultants. And in doing so, they confirmed something the consulting industry had been quietly hoping nobody would name directly.

The bottleneck was never the model. It was always the deployment.

What the Labs Are Admitting

OpenAI is embedding Forward Deployed Engineers directly within enterprises, working alongside business leaders and frontline teams to redesign workflows and build what they call durable systems. Anthropic is sending Applied AI engineers to sit next to clients and build custom solutions alongside them. Google is funding consulting firms to do the same work through the Gemini ecosystem.

Every one of these announcements is a version of the same confession. The model alone is insufficient. Getting AI from demonstration to production, connected to real data, governed appropriately, and trusted by the people who use it daily requires human expertise embedded close to the work. Software margins are attractive. Services revenue is how you actually get there.

Two of the most valuable AI companies in the world just spent $11 billion agreeing that AI deployment requires humans inside the business. Access to the API alone is insufficient.

What does that say about every company currently trying to manage their AI transformation with 100 ChatGPT licenses and a kickoff email?

The Commoditization Trap

For two years, the conventional wisdom in consulting was that AI would commoditize the deliverable but leave the relationship intact. A client who used to pay for analysis, synthesis, and recommendations can now get plausible versions of all three from a language model in minutes. But the consultant’s value lay in knowing which analyses mattered, synthesizing them in context, and building the trust that made recommendations actionable.

This turned out to be half right.

The deliverable is commoditized. A well-prompted language model produces a reasonable market analysis, a competent competitive landscape, and a plausible strategic framework in the time it takes to format a slide deck. The first draft of almost everything a junior or mid-level consultant produces is now effectively free.

What is left is the part that was always the actual value and was always underpriced because the commodity work provided the billable hours to support it. The judgment about which question to ask. The organizational context determines which recommendation is actionable versus theoretically sound. The relationship that makes the client willing to implement something difficult. The change management capability that turns a strategy into a behavior change.

The labs entering consulting are going after exactly this irreplaceable layer. Forward Deployed Engineers are selling embedded judgment and implementation capability. The model provides the commodity. The human provides the irreplaceable part. The question is who employs that human and what their relationship to the client looks like.

Mentormorphosis and the Trust Transfer

In Unmapping Customer Journeys, I describe a process called mentormorphosis. The gradual transformation of AI from a neutral assistant into a mentor-like figure in a customer’s mental hierarchy of trust and authority. The challenge for businesses is that clients trust AI more than they trust branded content, sales conversations, or even some consulting relationships. The commitment is reinforced through AI interactions more deeply than a client would get from traditional advisory engagements.

The AI labs entering consulting are accelerating this dynamic in a specific way.

When OpenAI’s Forward Deployed Engineer sits inside your organization, they are the bridge between the model and the business context. They bring the AI’s authority and the human’s judgment simultaneously. The client is buying deployment capability from the organization that built the model. They are buying deployment capability from the organization that built the model. The trust transfer is different in kind from what McKinsey or Deloitte has historically offered.

This is why the incumbents are responding. McKinsey is already cutting 10 percent of its workforce. Accenture, McKinsey, and Deloitte are simultaneously Google’s partners in the $750 million program and the targets of the labs’ direct deployment operations. The competitive position is genuinely uncomfortable. The consulting firms are being asked to help deploy the technology being used to replace their commodity work, while the technology companies are entering their premium relationship business.

The Deployment Gap Is Now the Most Valuable Real Estate in AI

The market is repricing the AI stack in real time.

The model layer is increasingly commoditized. Competition between labs, the availability of open-weight models, and rapid capability improvements across providers mean that any particular model’s advantage is temporary. The price per token has been falling consistently and will continue to fall.

The deployment layer holds the durable value. Connecting the model to real organizational data, making it work within actual workflows, governing it appropriately, earning the trust of the people who use it daily, and building the organizational capability to maintain it as the technology evolves. This is slow, human, contextual work that scales differently from software.

The labs are paying $11 billion to own this layer because they understand something important. The organization that controls the deployment relationship controls the customer. A client whose workflows are rebuilt around OpenAI’s deployment infrastructure, whose employees are trained on systems that OpenAI’s engineers designed, and whose governance structures reflect OpenAI’s implementation patterns is a client who will renew. The model contract is the downstream benefit of the deployment relationship.

This is vertical integration of the customer relationship, disguised as a services business.

What Survives in the Consulting Business

The consultants who will survive this transition are those with something the labs cannot replicate at scale.

Deep industry expertise that produces better questions rather than better answers. An AI model can synthesize existing industry knowledge. It cannot identify the specific organizational pathology that is preventing a particular client from implementing what they already know they should do. That diagnosis requires the kind of pattern recognition that comes from having seen the same failure in thirty different organizations over fifteen years.

Trusted relationships that predate the technology. The labs arrive in the deployment relationship as strangers. Capable and well-resourced, yes. Still strangers. A consulting relationship built over years, in which the client has shared information they would withhold from an AI model or an engineer from a technology company, is a different kind of asset.

Change management capability at an organizational scale. The bottleneck the labs identified is real. Getting AI from demo to production is hard because it requires people to change how they work. The technology is the easy part. The human systems are the constraint. The consultants with genuine capability in organizational change management, culture, and behavior change are doing something that the labs are still learning.

The consultants whose entire value sat in the analysis and synthesis that language models now produce are already in a different position than they were two years ago. The labs entering their business is confirmation. The cause was already in motion.

The Honest Assessment

The $11 billion the labs spent in 24 hours is, among other things, an honest answer to a question the consulting industry has been avoiding.

If the models are powerful enough to transform enterprise operations, why do enterprises need humans embedded alongside them to make it work?

Because transformation is a human problem. The technology creates the possibility. The humans make it real. The organizations that understand this earliest will build the capability to bridge the gap before the labs own that relationship entirely.

The ones that are still selling analysis and synthesis in a world where a language model does both in seconds have already lost the argument. The labs just made that official.

---

If you find this content valuable, please share it with your network.

Follow me for daily insights.

Book me to speak at your next event.

Start managing your agents for free.

Chris Hood is an AI strategist and author of the #1 Amazon Best Seller Infailible and Customer Transformation, and has been recognized as one of the Top 30 Global Gurus for Customer Experience. His latest book, Unmapping Customer Journeys, is available now!