TestMu AI adds self-repairing test engine to KaneAI platform

KaneAI's New Experience engine holds tests as objectives rather than click paths, letting an AI agent rewrite failing steps without losing proof of intent.

By The Datatech Times Editorial Team AI level 2 of 5
AI level 2 of 5: drafted by our AI editorial assistant from the company's announcement; fact-checked, edited and signed off by Mark Walker, Editorial Director. What the levels mean
A military-style unmanned ground vehicle with a central sensor turret stands on a dirt and grass testing ground with a pile of rocks and a small puddle in the foreground, and a red cube, blue tarp, marker poles, grassy hills, and distant tr

TestMu AI has shipped a new authoring engine inside its KaneAI product that reframes how automated UI tests are stored and maintained. Rather than recording a fixed sequence of clicks and selectors, the New Experience engine captures the intent behind each step as a discrete objective. When a product interface changes and a recorded step breaks, the engine can re-author the failing objective against whatever UI is currently in front of it, then hold that rewrite as a draft pending human review.

The company says the distinction matters because the standard repair technique in UI automation, patching broken CSS or XPath selectors, has a hard ceiling. When an interface is structurally reorganised rather than merely relabelled, no selector survives to patch, and only the underlying objective remains recoverable. TestMu AI's argument is that the authoring layer must change before maintenance behaviour can meaningfully improve.

What the engine does

New Experience ships with two operating modes under a single setting called Self-maintenance. The first, Adaptive Heal, re-authors the objective containing a failing step and all objectives that follow it, so a test run completes rather than halts mid-way. Objectives that already passed keep their recorded result. The second mode, Dynamic Test, goes further: it discards the recorded script entirely on every run and authors each objective fresh from intent, targeting interfaces that change so frequently that a static script becomes a liability before the next release cycle.

A complementary feature called Auto-approve determines whether a re-authored version immediately becomes the current test or waits for a human verdict. It ships disabled by default. Mayank Bhola, co-founder and head of products at TestMu AI, was direct about why: "A step can fail because a button moved or because the feature broke, and at the moment of failure those two look identical. Re-authoring around the second one turns a real defect into a passing test. The industry has sold self-healing as a way to stop looking at tests. We think it is a way to look at far fewer of them, far more carefully."

Re-authored versions appear in Version History attributed to the strategy that produced them rather than to the engineer who triggered the run. A reviewer can compare the new version against the current one and approve or decline; approving regenerates the exported code, while declining leaves the existing test untouched. An Evidence Pack report, covering test coverage and the business use cases behind a run, is available when every test case in a run uses New Experience and targets Chrome.

Market context

Automated UI test maintenance is a well-documented pain point in continuous delivery pipelines, and a growing number of vendors have positioned AI as the solution. Competitors including Mabl, Reflect and Functionize have offered AI-assisted selector healing and flow adaptation for several years, while established players such as Tricentis and Micro Focus have added similar capabilities to broader test-management suites.

The field is converging on agentic approaches: rather than reacting to a broken selector after the fact, newer tools attempt to reason about intent before the test run begins. TestMu AI's objective-level framing is consistent with that direction, though the company has not published independent benchmark data comparing repair success rates or false-negative suppression against named alternatives. Re-authoring in New Experience consumes authoring credits, introducing a usage-based cost dimension that enterprise buyers will need to factor into total cost of ownership alongside seat licences.

New Experience currently covers Desktop Web on Chrome, the same surface supported by the Kane CLI. TestMu AI has not announced a timeline for mobile or non-Chrome browser support, which will limit adoption in organisations running cross-browser or hybrid test suites. Buyers evaluating the platform should assess whether the credit model and the Chrome-only constraint align with their existing test infrastructure before committing to the new engine at scale.

Sector

Topic

Geography