Agentic AI · Quality engineering
Agentic AI across the product and testing lifecycle
Most teams use AI to write a bit of code. Mentarium built agentic solutions that run across the whole product and testing lifecycle, from the first user story to the release report, so the team's judgement goes into reviewing and deciding rather than producing every artefact by hand.
What the agents do
Plan
Draft features and user stories from the intent behind them, and analyse the quality of existing user stories: gaps, ambiguity and missing acceptance criteria.
Build
AI-first development of features, and migration of existing frameworks to current ones with the agent doing the mechanical translation and the engineer reviewing the result.
Test
Generate automation test scripts and manual test cases from the stories and the application, kept in step as both change.
Analyse
Impact analysis of pull requests, defect analysis, release analysis and preparation of engineering metrics, produced by agents and reviewed by the team.
What changed
The team spent less time producing stories, tests and reports and more time reviewing them, and the analysis that used to arrive too late to act on was available at each pull request and release.