Delegate, review, own: the operating model of AI-native teams
A model lab's playbook for AI-native engineering teams repeats one pattern across seven SDLC phases. The hard part is the review, where the engineer decides which agent findings matter.
Written by
-
Benedikt Stemmildt Founder & Co-CEO
Published: 28 May 2026 Updated: 2 min read
A major model lab published its playbook on building AI-native engineering teams. It walks through seven SDLC phases, and the same operating pattern shows up in every one of them.
Delegate, review, own.
The agent does the mechanical work. The engineer reviews the output. The engineer keeps ownership of the decision underneath it.
The pattern in each phase
Take planning: the agent generates the spec breakdown, the engineer owns the why. In testing, the agent writes the cases, and the engineer owns whether the coverage actually means anything. Code review is the sharp one. The agent flags the issues; the engineer owns which ones are real and which are noise.
Review means judging which findings matter
That last one is the move I think most teams haven’t internalized yet. The agent has plenty to say at code review. Some of it is right. Some of it is right in form but wrong in context. And some is just wrong. So the engineer’s job is to spot which of those issues actually matter for the system this code is going into. That is a different skill from spotting issues at all.
And I’d say it’s the skill the senior engineers I work with already have. They just hadn’t seen it named in this new shape.
Delegate does not mean let it run
The risk in adopting the playbook flat is reading “delegate” as “let it run.” Delegation in this pattern is closer to delegating to a junior engineer than handing off to a contractor. You set the context, you read the work, you stay accountable for what shipped.
That’s the same thing the modern software engineering crowd has been pointing at for a long time. Agents didn’t change the operating model. They just exposed whether your team had one to begin with.
The lab published the full seven-phase walkthrough. It is worth a read if you’re rolling out agentic engineering with your team.
In which SDLC phase has your team gotten the delegate-review-own balance right?
(Source: OpenAI, “Building AI-native engineering teams”, 2026.)
Further reading
Written by
-
Benedikt Stemmildt · Founder & Co-CEO
More on Organization and roles
- When the factory dispatches the work, the manager runs the line When the factory handles assignment and status, the engineering manager's job moves from dispatcher to line operator. Five responsibilities now define the role.
- Moving faster with AI only happens when it is a goal A software factory gives you output speed. Outcome speed needs the organization to change around it, and a change that isn't a goal loses to every change that is.
- Teams that rebuilt their org chart got more out of AI McKinsey surveyed about 300 enterprises on AI adoption. Top performers were 7x likelier to run AI-native workflows, while 70 percent have not changed roles at all. Pod restructuring is an upstream investment.
Scaling needs named owners
Coding agents scale across teams when named people own the platform, the team practice, the guardrails, and the measurement. Rollout & scale is where that ownership is built and handed to internal owners, in your stack.
Training on this topic: Agentic Engineering for Engineering and Product LeadersWant to check one team first? The self-check: twelve yes or no questions, three minutes, instant result, no account. Take the self-check