InnoWave has concluded AI4QA, an R&D project co-funded by Lisboa 2030 that applies generative AI to a recurring challenge in software quality assurance: deciding which tests to run first when there isn’t time to run them all.
The challenge
QA teams rarely lack test cases. What they lack is an objective, consistent way to decide which ones matter most in each release cycle. AI4QA, developed between January 2024 and June 2026, was designed to close that gap by optimising and automating test prioritisation.
The solution
The project delivered a platform built on three core elements:
- A modular six-component architecture that integrates with Jira, Xray and Office tools;
- A deterministic decision engine that assigns each test case a RiskScore that can be reproduced and audited;
- GAILEM, an AI agent that turns natural-language requests into prioritised, justified test plans, refined through conversation and automatically versioned.
At its core is a simple principle: the model explains, the engine decides. Generative AI makes prioritisation accessible and transparent, while every decision remains traceable.
The results
- 316 test cases derived from 54 user stories;
- 95.1% pass rate across executed test cases;
- A fully integrated platform, now in operation.
Watch the video
See how AI4QA works and what it means for QA teams.
AI4QA was co-funded by the Lisboa 2030 Regional Programme, Portugal 2030 and the European Union.
