← Developer Experience · Case studies Inner loop · Case study

AI-assisted code review & docs

An AI reviewer on every merge request — checking team standards, suggesting fixes, lifting unit-test coverage, and drafting documentation, callable from the IDE before a developer even pushes.

Overview

Reviews were a bottleneck: inconsistent feedback, long waits, and standards that lived in people's heads. The aim was to make the inner loop faster and more consistent by putting AI where developers already work — in the IDE and on every merge request — across the holistic product lifecycle (PLDC).

An automated reviewer triggers on each merge request, checks team standards, proposes fixes, and nudges unit-test coverage up; the same assistant drafts and refreshes documentation so it stays current.

The challenge

Code review was slow and a frequent source of delay
Feedback was inconsistent and standards lived in tribal knowledge
Documentation drifted out of date as code changed
Context-switching out of the IDE broke developer flow

Our approach

Deployed an AI code reviewer that triggers on each merge request — checking standards and proposing concrete fixes
Made it callable from the IDE before push, so issues are caught in the inner loop, not after
Used GenAI to draft and refresh documentation from the code and the change, keeping docs current
Wired coverage and quality signals into the pipeline so reviews lift unit-test coverage over time
Kept a human in the loop — AI accelerates and standardises review, it does not replace judgment

Results & business impact

Review feedback is faster and more consistent across teams
Standards are applied automatically rather than remembered
Documentation stays closer to current as code evolves
Developers stay in flow because review happens in the IDE

Tools & technology

AI code review GenAI in the IDE Documentation generation Unit-test coverage Merge requests Git CI/CD PLDC