A small team seen from behind standing at a wall calendar marked in three blocks of days, weeks and months, with a laptop on a table running a prototype, the whole scene drawn with wide empty margins on every side.

Thoughtworks

Quality is speed: Where engineering discipline moves in the AI era

Joe Murray with Martin Fowler · 6 October 2026

Company story · decisions · agents · After the prototype

Takeaway

Martin Fowler of Thoughtworks argues that teams using AI should move engineering discipline closer to working software instead of dropping it to gain speed.

Summary

Martin Fowler is chief scientist at Thoughtworks, a software consultancy. In an interview with his colleague Joe Murray, he discusses how teams can keep quality while AI agents speed up delivery. Murray presents the company's 3/3/3 framework: three days to shape an idea, three weeks for a working prototype and three months for a first product.

Key points

Fowler compares the present to the Extreme Programming method, which dropped heavy upfront design but moved rigour into automated tests and continuous integration.
He says each stage should end at a point where the team must stop and look at evidence, and that the exact durations matter less.
He argues that internal quality enables speed, and that code which is clear to people may also help AI agents work faster.
For AI agents with broad permissions, he says the discipline belongs in limiting what they can touch, restricting their network access and running endpoint protection.

Implication

The interview promotes Thoughtworks' own framework and agent platforms, and its client results are the company's own account. That a team can keep its speed by tying review points to working software rather than to plans is an inference and is not the authors' claim.

Plays

Engineering

Before an AI agent runs with broad permissions, restrict the files and systems it can change, limit its network access and run endpoint protection on its machine.

Product

End each stage of a project with working software that the team must inspect, not with a plan.

Derived by Working Surface from the article; more in the Playbook. Source line: Quality is an enabler of speed.

Source issue

8 October 2026: as AI agents produce code and prototypes faster than people can check them, teams must decide where human checks and decisions sit instead of letting review quietly lapse.