TestMu AI, the Full Stack Agentic AI Quality Engineering platform, announced New Experience, a new authoring engine in KaneAI that holds a test as set of objectives rather than recorded click path. Because objective survives change that recorded steps do not, KaneAI can re-author test against interface in front of it and hold rewrite as version for person to approve or decline.
KaneAI now authors on two engines. Original called Classic, new one New Experience, and every test case carries label showing which one authored it. New Experience currently covers Desktop Web browsers, same surface Kane CLI runs on.
Test maintenance has been standing cost of UI automation, and industry's answer has been to repair broken selectors. That answer has ceiling. When interface is restructured rather than renamed, there is no selector left to repair, and objective behind test is only part that survives. That is why authoring engine had to change before maintenance behaviour could.
Adaptive Heal re-authors objective containing failing step and every objective after it, so run finishes instead of stopping. Objectives that already ran keep result they replayed. Dynamic Test skips recorded steps and authors every objective from its intent on every run, for interfaces that change so often recorded script is liability.
Self-maintenance single setting holding both alongside Retry on Failure, with exactly one active at a time. Set for organization, project or run. Auto-approve changes decides whether re-authored version becomes current immediately or waits for verdict. It ships switched off.
Evidence Pack fuller report covering coverage and business use cases behind tests, available only when every test case in run is New Experience and on Chrome.
Re-authored version arrives in Version History as draft, attributed to strategy that produced it rather than to whoever started run. Reviewer compares it against current version and approves or declines: approving regenerates exported code, declining leaves current version untouched.
A test that repairs itself only useful if you can still say what it proves. New Experience holds test as objectives rather than recorded click path, so agent can rewrite path when product moves, because objective still there to rewrite it against. Industry has sold self-healing as way to stop looking at tests. We think it is way to look at far fewer of them, far more carefully.
New Experience available in KaneAI. Adaptive Heal and Dynamic Test apply to test cases authored with New Experience and configured for Chrome; every other test case replays recorded steps as before. Re-authoring consumes authoring credits, and run in which nothing fails to replay does not trigger it.
About TestMu AI
TestMu AI is full-stack agentic AI Quality Engineering platform that empowers teams to test intelligently and ship faster. Engineered for scale and AI-native by design, it offers end-to-end AI agents to plan, author, execute, and analyze software quality across web, mobile, and enterprise applications on real devices, real browsers, and custom real-world environments.
