> For the complete documentation index, see [llms.txt](https://docs.qapilot.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.qapilot.io/execution-and-reliability/ai-auto-healing-for-mobile-tests.md).

# AI Auto-Healing for Mobile Tests

## AI Auto-Healing for Mobile Tests

QApilot auto-healing attempts to recover a test step when its original locator fails after a minor UI change. It records healed steps in the execution report so you can review and approve locator updates.

#### **Overview**

QApilot’s **AI Auto-Healing** automatically repairs broken test steps caused by minor UI or element changes between app versions, ensuring your test cases remain stable and maintainable across releases. When a locator fails during execution, QApilot intelligently re-identifies the element using a multi-layered fallback process and updates the step context accordingly.

#### **How It Works**

**1. Locator Precedence**

QApilot resolves element changes in the following order of precedence:

1. **Element ID / Accessibility ID** – Primary and most reliable source.
2. **XPath & Tag Attributes** – Used for fuzzy matching when ID changes.
3. **Visual Match (Image Processing)** – Detects elements based on visual similarity when structural identifiers differ (e.g., color or layout change).
4. **Coordinate Fallback** – As a last resort, QApilot interacts with the element’s previously recorded screen coordinates.

If a step succeeds via a fallback method, QApilot marks it as **healed** in the execution log.

***

#### **Where to View Auto-Healing**

* After **test execution**, healed steps are highlighted in the **Execution Report**.
* Each healed step displays:
  * The **original locator**
  * The **AI-generated healing locator**

#### **Approving Healed Changes**

1. Navigate to the **Reports** section and open the relevant test execution.
2. Locate any step marked that has the "AI Assisted" tag as shown in the screenshot below
3. Click **“Update XPath”** beside the healed step.
4. The new AI-generated locator is automatically saved to the test case, and subsequent runs will use the updated version.

<figure><img src="/files/HWxfORVtKlYG2ZhUKCLr" alt=""><figcaption></figcaption></figure>

This ensures your test cases evolve automatically with your app, minimising manual maintenance.

For test-suite stability practices, see [Mobile Smoke, Regression, and Release Testing](/use-cases-and-best-practices/mobile-smoke-regression-and-release-testing.md).