> 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/ai-and-core-concepts/ai-native-and-agentic-mobile-testing.md).

# AI-Native and Agentic Mobile Testing

## AI-Native and Agentic Mobile Testing

QApilot applies AI to explore mobile apps, translate written test intent, and recover from selected UI changes. The platform combines autonomous and human-guided workflows with recorded test steps.

### What AI-native means in QApilot

AI is part of the testing workflow, not only a reporting add-on. QApilot uses it in documented capabilities such as:

* The Android crawler explores app paths and records discovered flows.
* CoWork plans actions from plain-English or BDD test cases.
* Auto-healing attempts alternate element identification when a locator fails.

These capabilities complement recorded tests. They do not remove the need to define expected outcomes and review results.

### Agentic workflows

An agentic workflow observes the current app state, selects an action, and can adapt when conditions differ. QApilot provides two documented authoring workflows:

#### Autonomous exploration

The [crawler](/ai-and-core-concepts/autonomous-android-test-generation.md) navigates Android applications with minimal manual input. It records interactions as test steps and stores results in crawl reports.

Use it to discover flows, build an initial coverage map, and identify issues during exploration.

#### Human-in-the-loop authoring

[CoWork](/ai-and-core-concepts/cowork-human-in-the-loop-mobile-test-authoring.md) starts from written test cases. It translates steps into editable BDD, acts on a live device, and replans after a failed action when possible.

A tester reviews the generated steps before accepting them. This keeps acceptance and release responsibility with the team.

{% hint style="info" %}
CoWork currently supports Android with LambdaTest. The crawler currently supports Android. iOS support for those workflows is documented as planned or in development.
{% endhint %}

### Application understanding and reusable coverage

During exploration, QApilot records screens, interaction paths, and test steps. The platform describes this structured output as a knowledge graph of app flows and screens.

Use that record as a starting point for reusable coverage. Promote important flows into test cases and suites. Add explicit assertions for business-critical outcomes.

### Adaptive recovery during execution

[AI auto-healing](/execution-and-reliability/ai-auto-healing-for-mobile-tests.md) uses locator fallback methods. The documented order includes IDs, attributes, visual matching, and coordinates.

A healed step is marked in the execution report. Review the suggested locator before updating the test case.

### When to use these workflows

* Use the crawler for early Android exploration and generated flow coverage.
* Use CoWork when existing written cases need mobile automation.
* Use auto-healing to reduce maintenance after minor UI changes.

### Related documentation

* [How QApilot Works](/ai-and-core-concepts/how-qapilot-works.md)
* [Run the Crawler](/ai-and-core-concepts/autonomous-android-test-generation/run-the-crawler.md)
* [AI Auto-Healing in QApilot](/execution-and-reliability/ai-auto-healing-for-mobile-tests.md)