Skip to content

Blog

Most AI rollouts change the codebase and leave the team untouched

In a Multitudes survey, codebase changes made up 85% of what teams changed for AI, and only 13% named a process change. In our experience, that gap is why most rollouts hit the same wall.

Written by

Adoption

Published: 12 May 2026 Updated: 1 min read

In our experience, most AI rollouts hit the same wall, often enough that we can predict it before the team can.

The codebase got smarter. The team did not.

A measurable pattern

That turns out to be a measurable pattern. In Multitudes’ survey (opens in a new tab), codebase changes made up 85% of the changes respondents made to work better with AI. Process changes, the most common change to how people work, came up with only 13% of respondents.

That matches our experience. The codebase is the easier surface. You can ship a CLAUDE.md by Friday. Adapting how reviews flow, how non-engineers contribute and how knowledge gets shared takes months and bruises egos.

What teams changed

The most common changes were markdown rule files (43%), documentation moved into the codebase (32%) and tooling integration (22%). All doable in an afternoon. None of them touch how engineers actually work together.

To protect quality, 31% of respondents brought in stricter human code reviews. More test coverage, a codebase change, came first at 39%.

The hygiene the field already knew

One principal engineer in the data said the work around Claude is “the things that we’ve been talking about for the last 3 years”: cutting down PR sizes, small chunks of work, test-driven development.

That is the same hygiene the field has been pointing at for a decade. AI just made the cost of skipping it visible.

Agentic engineering is more than a codebase question. It is how the codebase, the engineers, the reviews and the non-engineers move together. Most companies have only changed the first one.

What did your team actually change, and what didn’t you?

(Source: Multitudes whitepaper, March 2026)

Further reading

  1. Agentic adoption is harder than agile, because you can't fake it Agile rollouts could rename roles and keep the waterfall. Agentic engineering returns nothing on half-measures, because every gate sits upstream in architecture, context and the team. 2 min readShort post
  2. Agency, not the tool, decides whether AI works for a team Skill on one axis, agency on the other: four quadrants with four very different outcomes. Give senior developers agency and the tool starts working. Strip it away and even the best reject it. 2 min readShort post
  3. AI makes teams faster and wears down their engineering instincts Agents speed up output while problem decomposition, debugging and code comprehension fade. The answer is to encode the team's engineering knowledge into a harness. 2 min readShort post

All 13 posts on Adoption

ClientsCase study: Tegtmeier Internet Solutions

Test the constraint in your system

The Factory Readiness Assessment connects this research to your value stream, your data, and the next decision.

Training on this topic: Agentic Engineering for Engineering and Product Leaders
See the assessment

Want to check one team first? The self-check: twelve yes or no questions, three minutes, instant result, no account. Take the self-check