MaxTAF 2.8.0: say what you want tested, and get a regression test back

MaxTAF 2.8.0 records what the AI does while it tests, and hands it back as an ordinary MXML case. The thinking is paid for once, when the test is built; what runs every night afterwards is a classical script.

Until now you chose between two trades. A hand-scripted case is deterministic and costs almost nothing to run, but somebody has to write it and keep it working. An AI case takes minutes to create in plain English, but the AI does the work again on every run, so each run costs money and no two take quite the same route. In 2.8.0 you describe the test once, the AI carries it out against your live Maximo, and you are left with a case in your Workspace, in the same language your existing suite uses. Read it, press Compile once, and it runs on demand or on a schedule. We switch recording on for your project; nothing changes on your side.

The AI Assistant changed too. It reads a failed run, proposes a fix, applies it, recompiles and re-runs so you see whether the case now passes. Every change it makes is listed in the Change Timeline with one-click undo, and it runs on your own provider keys.

Each feature has an article of its own: Build a test once with AI, then run it as MXML and Improvements to the AI Assistant, both on the release page.