Stage 3: Build

Deliver working software in small, reviewed steps, using AI to help and with people responsible for the result.

Goal: Deliver working software in small, reviewed increments.

Each story meets the definition of done; the increment is demonstrated to the customer.

How a one-week sprint works

A one-week sprint: with AI-assisted engineering we plan, build and release every week. Length and days can be adapted per engagement.

1
Pick up a ready story
The engineer takes the next story that meets the definition of ready and confirms understanding with the BA if needed.
2
Implement with AI assistance
AI tools may draft code and tests under the AI-use policy. The engineer understands every line.
3
Test and self-review
Automated tests are written or updated; the engineer reviews their own diff before asking others.
4
Peer review
Another engineer reviews. Changes touching architecture, security or data are reviewed by the Tech Lead or Security.
5
Merge through the pipeline
CI runs tests, static analysis and dependency checks. Failing checks block the merge.
6
Demo and release
At the end of the one-week sprint the team shows working software to the customer, collects feedback and releases what has passed the gates.

Each stage ends with an exit gate agreed with the customer. What we learn in Operate feeds the next round of Discover.

Lifecycle

Prove that the software works, is secure and can be released.

Quality & AI

Four gates a change passes between idea and production, and what each one checks.

Quality & AI

Where we use AI, the rules that keep it safe, and who is responsible.