Skip to content

AI Analyzer — Automatic Failure Analysis

When a Test Suite's success rate drops between executions, the AI Analyzer helps you understand what went wrong. It investigates the failed run, explains the likely root cause in plain language, and often suggests how to fix it.

The AI icon on the Executions page

A brain icon in the Logs column of the Executions page marks a run that has, or can get, an AI failure analysis. It comes in two colours:

  • Orange, "View failure analysis": automatic. When a finished suite's success rate drops compared with its previous finished run (for example, from 100% down to 0%), the analysis runs by itself and the icon appears next to the affected execution, marking the run where the regression happened. Re-runs and runs stopped by a user are not compared.
  • Purple, "Run for failure analysis": on demand. On environments where on-demand analysis is enabled, any finished run with failures or initialisation errors shows this icon. Click it and then Analyze to start an analysis.

Executions list with the AI icon on the regressed run row

Running an analysis

Click the icon to open the Failure analysis — window. If there is no analysis yet, click Analyze: the message "Analysis started — analyzing…" confirms it, and the results appear in the window (and in the list) once the analysis completes, usually within a few minutes. If an analysis failed, you can start it again.

The AI Analyzer investigates why the drop occurred. It reviews the failed steps and recent run history, then explains the likely root cause in plain language — and often suggests how to fix it.

Failure analysis modal listing the analyzed failures

The results show how many failures were analysed, an Across the run summary with the confidence and any Shared cause, and one card per failure (for example TC 3/18: ). When the environment has no analysis available, the window reads "No failure analysis is available for this execution."

The same analysis is shown per Test Case in the Show Logs window.

What it analyzes

  • Failed step details
  • Recent run history
  • Configuration or test-case changes made shortly before the run

Note

The orange icon only appears when a measurable drop in success rate is detected between executions, so it points you directly to the run where the regression began. Whether the purple on-demand icon is available depends on your environment; ask your Quadified contact.