AI Agents

Autonomous AI testing

QA.tech uses specialized AI agents to test your application like humans would, but faster and more thoroughly. Our system automatically routes testing tasks to the right agent based on context-no manual configuration needed.

Our AI Agent System

Entry Point	Purpose	How It Activates	Learn More
Chat Assistant	Interactive testing, test creation, site analysis	You start a conversation	AI Chat Assistant
PR Review	Autonomous testing of every pull request	Automatic on PR open/update	GitHub App
On-Demand PR Testing	Deep testing with custom instructions	@qatech mention in PR comments	GitHub Actions

How It Works in Practice

Here’s what happens when you open a PR that changes your checkout flow:

1. Automatic Detection

PR opened: "Add Apple Pay to checkout"
→ PR Review agent activates

2. Change Classification

Analyzing diff...
├─ checkout.tsx: USER-FACING ✓
├─ payment-methods.ts: USER-FACING ✓
└─ README.md: DOCS ONLY ✗

Classification: USER-FACING changes detected

3. Coverage Assessment

Existing tests found:
├─ "Complete checkout with credit card" ✓
├─ "Complete checkout with PayPal" ✓
└─ Apple Pay integration: NOT COVERED ⚠️

4. Test Generation

Creating new test:
"Complete checkout flow with Apple Pay"
├─ Add item to cart
├─ Navigate to checkout
├─ Select Apple Pay as payment method
├─ Verify payment confirmation
└─ Verify order appears in account

5. Execution & Review

Running against PR preview environment...
✅ All tests passing (3/3)

Posting GitHub review:
"✅ Tests passing - Apple Pay integration verified"

This entire workflow runs autonomously - no human intervention required.

When to Use AI Agents vs Scripts

Scenario	AI Agents	Scripts	Manual QA
Exploratory testing	✅ Best - discovers edge cases through varied behavior	❌ Fixed paths only	✅ Good - but slow and expensive
Regression testing	✅ Good - handles UI changes gracefully	✅ Best - fastest execution	❌ Too repetitive
Exact same steps every time	⚠️ Possible but overkill	✅ Best - fully deterministic	❌ Error-prone over time
Testing multiple user flows	✅ Best - tries different approaches automatically	⚠️ Need separate script per variant	✅ Good - but doesn’t scale
Complex calculations	❌ Use assertions in code instead	✅ Best - precise math	⚠️ Manual calculation prone to errors
Form filling with realistic data	✅ Best - understands context	⚠️ Need hardcoded test data	✅ Good - but tedious
Testing preview environments	✅ Best - automatic PR reviews	⚠️ Need CI/CD integration	❌ Requires manual coordination
First-time test coverage	✅ Best - AI Chat Assistant sets up quickly	❌ Requires upfront investment	⚠️ Slow to establish baseline

In practice, most teams use a combination: AI agents for exploration and edge cases, scripts for critical deterministic flows, and manual QA for subjective evaluation (design, UX, brand consistency).

How It All Connects

The system coordinates automatically based on what you’re trying to accomplish:

Creating tests via Chat

You: "Create 5 tests for the checkout flow"
→ Analyzes your application structure
→ Creates tests prioritizing revenue-critical paths
→ Shows suggestions for your approval

PR Review fills coverage gaps

PR detected with untested functionality
→ Creates tests for new features
→ Runs all relevant tests
→ Posts review with results

You don’t manage this coordination - it happens automatically.

Test Execution Model

QA.tech uses Claude Haiku 4.5 as the default AI model for test execution. This model provides the fastest test execution while maintaining high-quality results.

Model	Speed	Best For
Claude Haiku 4.5 (default)	Fastest	Most tests - recommended for day-to-day testing
Claude Sonnet 4.5	Moderate	Complex scenarios requiring deeper reasoning

The AI model handles all test execution decisions: navigating your application, filling forms, clicking buttons, and verifying outcomes. You write test goals in natural language, and the model figures out how to achieve them.

You can override the agent per-test in the Advanced settings when creating or editing a test. See Creating Tests for details.

Get Started

Create Your First Test

Write tests in natural language. The AI generates steps automatically.

AI Chat Assistant

Conversational interface to create, edit, and manage tests.

Automatic PR Testing

Install the GitHub App for autonomous reviews on every PR.

CI/CD Integration

Trigger tests from your deployment pipeline.