At Your Service: Concierge-style AI Troubleshooting

Introducing Leapwork CoPilot Investigation

In enterprise testing, time spent diagnosing failed test runs can quickly become time lost, especially when systems are complex, cross-platform, and are constantly evolving. Quality Engineers often find themselves sifting through logs, reviewing screenshots, and retracing steps just to uncover a root cause.

Enter Leapwork CoPilot Investigation Analysis: a new AI-powered assistant designed to help teams troubleshoot test failures faster and more consistently.

The challenge: Test failures without clear explanations

Imagine you're a QA lead using traditional tools or manual testing. You're responsible for validating an end-to-end order flow across a retail system. A test fails in your nightly regression suite. There's no obvious error in the log. You suspect it's a timing issue with a dynamic UI element—but confirming that requires log inspection, screen review, and possibly even re-running the test with debug logs enabled. Multiply this by dozens of test cases across distributed teams, and troubleshooting becomes a costly bottleneck.

This is exactly the type of scenario Leapwork CoPilot Investigation was built to address.

Two ways to use Leapwork CoPilot Investigation: On-demand or always-on

1. On-demand investigation (default setting)

This option is perfect for testers who prioritize fast execution and want to use AI assistance only when needed. When a test flow fails, users can manually trigger the investigation by clicking “Get CoPilot Investigation Analysis” in the run log preview.

Once triggered, Leapwork sends execution keyframes and error context to the Leapwork CoPilot, which returns a short, structured root cause analysis, including:

• Root Cause: A set of one or more suggested root causes (e.g., element not found in time in the DOM)
• Recommendation: An actionable next step (e.g., use AwaitDOM change on the block that caused the failure)

The result is added as a new log entry labeled “Experimental,” keeping the rest of your workflow unchanged. It's fast, helpful, and doesn't require advanced configuration.

2. Automated investigation

For teams that want consistent AI-driven diagnostics across every failed run, Leapwork CoPilot Investigation can be enabled to run on every failed flow globally via Default Flow Settings. Once toggled on, every failed execution automatically triggers an AI analysis, with insights logged directly into the run report.

This mode is ideal for QA teams that operate at scale and want Leapwork CoPilot Investigation to help accelerate troubleshooting with potential root causes into why a test failed—without manual review or guesswork. Think of it as having a built-in AI debugger for every failure, working 24/7 across all environments.

Designed for adoption without disruption

Leapwork CoPilot Investigation is intentionally low-friction. Manual mode ensures testers can try it without impacting current processes, and the “Experimental” label signals that feedback is welcome as the feature evolves. Documentation is readily available, and the feature rollout is designed to be gradual, ensuring users discover its value organically.

Final thoughts

Leapwork CoPilot Investigation is not just about AI. It's about making QA teams more effective, reducing guesswork, accelerating troubleshooting, and strengthening the overall feedback loop between failure and fix. For Quality Engineers navigating increasingly complex systems, this is the kind of intelligent assistance that makes a tangible difference.

Try it in your next test run, or enable it globally to see the difference at scale.

👉 Learn about Microsoft Copilot automated testing with Leapwork