The enablement gap: labs now sell the build, not the model (Ode + Deployment Company). "You can buy the build, not the fluency." News-pegged.

Anthropic’s $1.5bn implementation venture now has a name: Ode.

The more interesting fact is that it exists at all.

Within a week in May, Anthropic and OpenAI both announced separate businesses dedicated not to building better models, but to making those models work inside companies. Anthropic has Ode. OpenAI has its Deployment Company.

That is a fairly expensive admission: model capability does not automatically become business value.

Their answer is forward-deployed engineers. Elite technical teams, often former founders, embedded inside customers to redesign workflows and build bespoke systems around them.

Ode’s chief technologist put it well. Model selection matters, but it is only one ingredient in a system that still has to be engineered.

He is right. But look at what Ode selects for, because the missing ingredient is hiding in the customer criteria.

Its ideal customer is not simply a company with a difficult technical problem. It is one where the CEO has already bought in and the work ranks among the organisation’s top priorities.

That is not an engineering requirement. It is also not the whole organisational requirement.

Executive sponsorship buys budget, permission and attention. It does not create people who know how to judge the system’s output, adapt it to changing conditions or notice when it is quietly wrong.

That is the gap the implementation model still does not solve. Most organisations have developed very few people with that fluency. Give them a beautifully engineered workflow and they may use a fraction of it, work around the rest and gradually return to what they know.

Technically impressive, yet lightly used and expensively shelved.

The same problem runs in the other direction, into the choice of use case. The problems worth solving are often understood by the people closest to the work, not by the executive signing the contract or the engineer arriving to build. A system specified two layers away from the problem can be flawless and still be aimed at the wrong thing.

The labs have correctly moved beyond selling access to a model. They now want to sell the build around it.

But a build is only capability. Value appears when people equipped to use it apply it to a problem someone genuinely has.

You can buy the build, but you still have to build the fluency to use it.