# How AI test automation works

> Hand testing to Cue April: connect your repository, approve a plan and let an AI quality lead coordinate specialist checks and bring back findings.

Source: https://cueapril.com/how-it-works
Updated: 2026-10-12
Index: https://cueapril.com/llms.txt

## Step 1: Set the mission

**Tell your testing lead what good looks like.** Connect GitHub or Azure DevOps, choose the exact source revision and bring your agreed requirements. Your lead discovers supported existing suites and proposes a setup around the change that matters to your team.

## Step 2: Approve the plan

**Set the boundaries. Hand over the legwork.** Review the methods, targets, environment actions and run allowance. Your quality lead turns your quality bar into a focused investigation. Testing starts after the required setup and approval; creating an account does not authorise a run.

## Step 3: Put testing to work

**One lead. A coordinated testing effort.** Your lead coordinates approved tasks across the methods your change needs: existing suites, API and browser checks, generated inputs and mutation testing. Tasks use your configured environment and agreed expectations. New assumptions come back for review.

## Step 4: Act on the findings

**Get findings, with the evidence attached.** See the tested revision, expected behaviour, actual outcomes and retained artifacts together. Passing, failing, blocked and incomplete work stay visible. Reproduce supported findings, review regression drafts and run fresh checks to verify a fix.

## Step 5: Keep the momentum

**Keep testing as your software moves.** Schedule approved checks or watch selected branches to repeat testing within your standing scope and allowance. Each run needs current authority, an available runner and enough budget. Your team reviews the evidence, resolves the gaps and decides what ships.
