Adoption is the second sale.

The deal closes. Everyone posts about it. Nine months later the account comes up for renewal and somebody discovers that forty licences were provisioned, eleven were ever used, and the executive who sponsored the purchase has moved to another company.

That story is common enough to be boring, and it is usually filed under customer success. It belongs in go-to-market, because the conditions that produced it were set during the sale.

The renewal wave is the deadline

There is a specific reason this matters more in 2026 than it did in 2024. The first substantial cohort of enterprise AI purchases is reaching its second and third renewal, and the conversations are different from the first ones.

Bessemer put the timing plainly in its pricing and monetisation playbook: as 2025 pilots hit 2026 renewals, pricing has to reflect actual value rather than promise. Their related observation is the harder one. Soft return-on-investment positioning kills willingness to pay, and the example they use is the copilot that gives advice without closing the loop, leaving the customer asking whether they are really getting value.

At the first renewal, enthusiasm can carry a deal. At the second, somebody produces a usage report.

The evidence on where value gets lost

The most quoted number in enterprise AI is MIT’s finding that only about 5% of generative AI pilots delivered rapid revenue impact. The reason matters more than the figure, and the authors attributed it to a learning gap in both tools and organisations rather than to model quality, noting that generic tools work for individuals and then stall inside enterprises because they do not adapt to specific workflows.

Forrester reached a similar conclusion from a different direction in April 2026, reporting from a survey of 1,500 AI decision-makers that most enterprises are struggling to convert rising adoption and investment into measurable business impact. One of the barriers it names should make every vendor uncomfortable: an overemphasis on productivity-focused use cases. Selling time savings is easy and it produces exactly the kind of value nobody can find at renewal.

Both findings say the same thing. The failure happens after the purchase, in the space between a licence and a habit.

There is a measured version of the same problem inside a regulated sector, which is worth reading as a leading indicator. The Bank of England and FCA found that 46% of firms had only a partial understanding of the AI technologies they use, against 34% claiming complete understanding, and attributed the gap largely to bought rather than built systems. People do not adopt what they do not understand.

What go-to-market owns after the signature

Four things, and none of them belongs to customer success alone.

The success definition, written before the contract. What number moves, by when, measured how, compared against what. If that sentence does not exist in writing at signature, the renewal conversation becomes an argument about impressions. Sales resists this because it introduces a target that can be missed. That is the point.

A first-90-days plan the customer can see. Which of their roles has to exist, what has to be configured, what training happens when. Publish it publicly before the sale rather than delivering it after, because it is also a differentiator: most competitors will not show theirs.

Internal enablement material for the customer’s own teams. Your buyer has to sell this internally to the people who will actually use it, and they are not equipped to do that. Give them the explanation, in language a sceptical team member will accept.

A usage signal you both watch. One number, visible to both sides, reviewed monthly. Not a dashboard with forty metrics. A single figure that answers whether this is being used for the thing it was bought for.

The measurement that gets you the renewal

Gartner’s research contains an uncomfortable finding for anyone who has optimised for self-service. 75% of B2B buyers say they prefer a rep-free experience, and self-service digital purchases are considerably more likely to end in purchase regret. The same research puts buyers at 1.8 times more likely to complete a high-quality deal when they use supplier-provided digital tools alongside a rep rather than working through it alone.

Regret is a renewal problem wearing a sales costume. A buyer who bought easily and then failed to get value does not blame their own process.

Which suggests something counter to the prevailing direction of travel. The right amount of friction, in the form of a human establishing what success requires before money changes hands, protects the second sale. Removing every obstacle to purchase optimises the metric you can see and damages the one you cannot.

The adoption story is also your best marketing asset

The commercial argument for taking adoption seriously is not just retention.

In a market where the dominant belief is that AI deployments fail, a customer who can describe a measured result in their own words is the most valuable asset you own. It survives procurement review, it gets corroborated externally, and it is the thing a language model retrieves when somebody asks whether your product works.

You cannot manufacture that from a launch. It comes from having designed the deployment to produce a number, and from having asked for the number at the right moment, which is roughly month four, once the result exists and before the sponsor changes jobs.

The check

Take your ten largest accounts. For each one, write down the number that account was bought to move, and where that number stands now.

If you cannot fill in the second column for more than half of them, your renewals are being decided by people with better data than you have.