Test Analytics Agent

Ask queries on your tests, get AI-generated summaries on dashboards, and view insights on your test suites.

The Test Analytics Agent provides insights from test data. It explains why the data points in your tests are the way they are, recommends fixes, and links insights across projects.

Functions performed by the agent

The Test Analytics Agent organizes its capabilities into three main functions:

The functions work together:

• Generation creates the view you need.
• AI-generated insights highlight changes that matter.
• The agent answers your follow-up questions, explains causes, and supplies an evidence-based action plan.

Enable BrowserStack AI

You must enable BrowserStack AI before using the Test Analytics Agent. You need the Owner or Admin role to change the BrowserStack AI settings.

You can enable BrowserStack AI in your account settings.

Ask questions about your test data

Type your question in plain language. The agent scopes its answer to the context and time window of the page you are on.

The agent covers the following use cases:

The agent answers questions in five intent dimensions:

• Facts: the current state.
• Trend: the change over time.
• Breakdown: results segmented by OS, browser, or test.
• Diagnose: the root cause.
• Prescribe: what to fix, and in what order.

The agent answers questions in the following categories:

• Answers the why along with the what. The agent explains the root cause and suggests what to fix, rather than just providing a failure rate.
• Fewer steps. Simplifies the manual multi-screen analysis into one answer.
• Turns failures into a fix plan. The agent groups failed tests across projects into a few root causes, ranks them, and adds fix guidance.
• Connects causes across projects. Identifies when one root cause affects multiple projects.
• Runs Test Failure Analysis on confirmation. When the root cause hasn’t been analyzed yet, the agent can trigger Test Failure Analysis after you confirm.

Turn dashboard data into insights

You can ask questions from the side panel on any dashboard page:

1. Navigate to a dashboard page.
2. Open the Ask AI panel.
3. Type your question in plain language. For example: “What’s the flakiness trend for the login workflow over the last 14 days?” or “Is this build safe to ship?”
4. Review the answer. It includes links to the relevant views and to the evidence for each conclusion.
5. Ask follow-up questions, such as “What should I fix first?” or “Why is that test flaky?”

Get insights from reports

You can ask questions on Build Comparison reports and Test Execution reports. Open a report page and use the Ask AI panel to ask about the results. For example: “Which tests failed in this report and why?” or “How does this build compare to the previous one?”

The agent answers from the data in the open report. Answers can include failure breakdowns, pass rate trends, and build-over-build comparisons.

Get answers on project-specific pages

The agent answers questions scoped to a single project when you ask from a project-specific page. The same intent dimensions (Facts, Trend, Breakdown, Diagnose, Prescribe) apply, specific to that project’s data.

Get answers on multi-project pages

On multi-project-level pages such as Overview and Reports, the agent answers questions across your entire test suite, including cross-project root cause analysis.

Get actionable insights from any widget

The Test Analytics Agent delivers plain-language summaries on your dashboards, Build Runs, and Unique Errors pages when users ask for one.

You can follow up on any summary with a question. When you see an insight and want more details, ask “Why?” and the agent continues from there.

The following pages show insights:

Summaries follow these rules:

• The agent never restates a number the dashboard already shows. Every summary explains what the number means.
• The agent informs only when something has changed.
• Summaries reflect the latest data and update when data changes.
• Every summary is designed to prompt a follow-up question you can ask the agent.

Generation - create dashboards and reports by asking in plain language

Generation helps you create and edit dashboards, widgets, and reports by asking in plain language.

The following capabilities are available:

• Create dashboards using plain language: Build or edit dashboards, widgets, and filters by describing what you want. For example: “Create a dashboard to visualize the failure rate by device for the login action in the last 3 months” or “Create a widget to track flaky test count in each project.”

• Edit existing dashboards: Adjust filters, modify widgets, or change layouts by asking in plain language. For example: “Modify the time range on this widget to 90 days” or “Filter this dashboard to show only the checkout and payment screens.”

• Create reports using plain language: Build Test Execution reports or Build Comparison reports by asking in plain language. For example: “Compare the latest two builds for the checkout project” or “Generate a test execution report for the last sprint across all projects.”