HOW AI TESTS SOFTWARETHE COURSE / JASON ARBON

Parts One–Six recorded · Free previews

Become an
AI-first engineer.

A fast, practical path to better testing with AI.
Find bugs, build tests, investigate risks, and make better release decisions—with skills you can use on your next project.

The full course is live on Udemy.

Six hours of practical AI testing. Start learning today.

Enroll on Udemy ↗
Explore the course

With Jason ArbonUpgrade how you work as a tester, developer, or QA leader.

YOUR NEXT SKILL UPGRADE STARTS HERE.
Jason, with a shaved head and beard, teaches a class. Play the course overview. AI CODING AGENT > Test this coffee shop. Explore the menu… Try a vanilla latte. Check cart and price… Compare the receipt. Save the evidence. Cart verified Discount mismatch ILLUSTRATIVE RUN TestBucks Your coffee. Your way. Order a coffee PLAY OVERVIEW Indian student with short dark hair, seen from behindEast Asian student with a ponytail, seen from behindWhite student with a shaved head, seen from behindDark-skinned student with an afro, seen from behind

Supported

Your tools. Your stack. The same testing principles.

AI assistantsClaudeChatGPTCodexCursorGitHub Copilot
AutomationPlaywrightSeleniumVibiumCypressAppium
Issue trackingJiraGitHub IssuesGitLab IssuesLinearYouTrack
Test managementTestRailXrayZephyrqTestTestmo
LanguagesPythonTypeScriptJavaJavaScriptC#
Test cloudsBrowserStackSauce LabsLambdaTestAWS Device FarmFirebase Test Lab
PlatformsiOSAndroidWindowsmacOSLinux
BrowsersChromeFirefoxSafariEdgeChromium
InterfacesMCPREST APIsGraphQLOpenAPIWebDriver BiDi

Apply the course methods across these ecosystems. Available actions depend on your agent’s tools, connectors, permissions, and setup—not built-in integration with every product. Chat-only tools cannot operate a browser without browser access.

View all 45 tools & technologies

AI assistants

ClaudeChatGPTCodexCursorGitHub Copilot

Automation

PlaywrightSeleniumVibiumCypressAppium

Issue tracking

JiraGitHub IssuesGitLab IssuesLinearYouTrack

Test management

TestRailXrayZephyrqTestTestmo

Languages

PythonTypeScriptJavaJavaScriptC#

Test clouds

BrowserStackSauce LabsLambdaTestAWS Device FarmFirebase Test Lab

Platforms

iOSAndroidWindowsmacOSLinux

Browsers

ChromeFirefoxSafariEdgeChromium

Interfaces

MCPREST APIsGraphQLOpenAPIWebDriver BiDi
AllWorks in all AI coding agents
73Lessons
6Hours
9Previews
WHAT YOU’LL BE ABLE TO DO

Make AI part of
your testing advantage.

Improve the way you test—from your first bug investigation to automated checks and release confidence. Practice on TestBucks, then bring the methods to your own product.

  1. 01

    Find bugs you would otherwise miss

    Use AI to investigate screens, logs, code, and data. Turn more evidence into useful findings, without treating every AI answer as fact.

  2. 02

    Build tests that adapt with your product

    Move from one-off prompts to reusable tests and workflows. Apply AI across functionality, security, accessibility, performance, and more.

  3. 03

    Make stronger quality decisions

    Prioritize the risks that matter, challenge weak evidence, and explain whether a release is ready. Put your new skills into practice.

Jason Arbon, your course instructorYOUR INSTRUCTOR
ALMOST 20 YEARS. SHARED WITH YOU.

Learn with Jason Arbon.

Almost 20 years focused on testing AI—and using AI to test software.

Jason uses AI across every aspect of testing—from analyzing risks and investigating bugs to building and running tests, evaluating results, and assessing release confidence. Learn his methods and put them to work on your own software.

These workshops have been refined through presentations at several of the world’s largest testing and AI conferences.

“My personal mission is to improve the quality of the world's apps.”

Some of the lessons in this course also come from Jason’s experience raising $30 million from Google to focus on testing and AI.

COMPANIES & PRODUCTS

Built on experience.

Google
Google Search
Chrome
Microsoft
Bing
Windows
test.ai
Testers.ai
IcebergQA

Companies and products Jason has worked on or with.
Experience, not endorsements.

BOOKS BY JASON

Four books. One testing obsession.

How Google Tests Software coverHow Google Tests Software ↗How AI Tests Software coverHow AI Tests Software ↗Testing AI coverTesting AI ↗App Quality coverApp Quality ↗
TRY A SAMPLE. SEE THE VALUE.

Put the lessons to work.

Use practical, copy-and-paste prompts in your AI agent to apply the course methods to your own testing. Start with three worked exercises, two prompt samples, and two screenshot examples. Unlock the complete collection with your course access code.

Meet TestBucks

Use our fictional coffee shop to practice with your own AI coding agent.

Open TestBucks ↗
AVAILABLE NOW ON UDEMY

Put AI to work in your testing.

Your next skill upgrade starts here.

Enroll in the full How AI Tests Software course and learn at your own pace on Udemy.

  • Six hours across six course parts
  • Practical demonstrations and hands-on exercises
  • Methods you can apply to your next project
Enroll on Udemy ↗

View current pricing and purchase the course on Udemy.

Watch free previews →

SIX PARTS. ONE PRACTICAL PATH.

Find your next lesson.

73 lessons · 6 hours
Nine free previews. No sign-in needed.

In color: watch a free previewGrayscale: full-course video only

AI Checks

22 lessons · 90 min
3:41
VIDEO 01 · FREE PREVIEW

Welcome to How AI Tests Software

Who the course is for, what you will learn, and where the six-part journey is headed.

IN THIS VIDEO

Frame from Set up your AI coding agent4:25
VIDEO 02 · FULL COURSE

Set up your AI coding agent

Get an agent reading local files, set safe boundaries, and choose a suitable practice environment.

IN THIS VIDEO

  • Lesson 02: Set up your AI coding agent Full course · locked
Frame from Meet TestBucks. Find your first bugs.1:49
VIDEO 03 · FULL COURSE

Meet TestBucks. Find your first bugs.

Use a fictional coffee shop to make your first evidence-backed observations.

IN THIS VIDEO

  • Lesson 03: Meet TestBucks. Find your first bugs. Full course · locked
Frame from Checks, tests, harnesses, and confidence5:35
VIDEO 04 · FULL COURSE

Checks, tests, harnesses, and confidence

Separate artifact inspection from interactive testing and longer-running quality workflows.

IN THIS VIDEO

  • Lesson 04: Checks, tests, harnesses, and confidence Full course · locked
4:54
VIDEO 05 · FREE PREVIEW

AI checks: start with the evidence

Inspect screenshots, logs, and source artifacts. Learn what a screenshot can and cannot prove.

IN THIS VIDEO

Try: Find bugs in one homepage ↗
Frame from Write a bug report that helps fix the bug4:56
VIDEO 06 · FULL COURSE

Write a bug report that helps fix the bug

Use OpenTest.ai-style fields for reasoning, counterarguments, routing, and a fix prompt.

IN THIS VIDEO

  • Lesson 06: Write a bug report that helps fix the bug Full course · locked
Frame from Find issues in logs and recorded evidence6:06
VIDEO 07 · FULL COURSE

Find issues in logs and recorded evidence

Inspect network traces, console logs, and database artifacts, not just the visible page.

IN THIS VIDEO

  • Lesson 07: Find issues in logs and recorded evidence Full course · locked
Frame from Connect findings to your existing tools2:20
VIDEO 08 · FULL COURSE

Connect findings to your existing tools

Connect issue and test repositories while keeping a human review step before filing.

IN THIS VIDEO

  • Lesson 08: Connect findings to your existing tools Full course · locked
Frame from Give your agent useful context5:15
VIDEO 09 · FULL COURSE

Give your agent useful context

Combine product rules, code, logs, and team knowledge to investigate more effectively.

IN THIS VIDEO

  • Lesson 09: Give your agent useful context Full course · locked
Frame from Many pages. Fewer duplicate reports.6:47
VIDEO 10 · FULL COURSE

Many pages. Fewer duplicate reports.

Find inconsistencies across pages and consolidate repeated findings into actionable issues.

IN THIS VIDEO

  • Lesson 10: Many pages. Fewer duplicate reports. Full course · locked
2:36
VIDEO 11 · FREE PREVIEW

Ask focused questions

Turn known failure modes into reusable checking questions instead of one-off hard-coded cases.

IN THIS VIDEO

Frame from When the assistant refuses5:03
VIDEO 12 · FULL COURSE

When the assistant refuses

Understand refusals, authorized testing context, and the product behavior that follows a refusal.

IN THIS VIDEO

  • Lesson 12: When the assistant refuses Full course · locked
Frame from Security and privacy checks1:58
VIDEO 13 · FULL COURSE

Security and privacy checks

Look for cross-customer data, unintended disclosures, and inconsistencies in captured evidence.

IN THIS VIDEO

  • Lesson 13: Security and privacy checks Full course · locked
Frame from Usability and accessibility4:14
VIDEO 14 · FULL COURSE

Usability and accessibility

Examine confusing choices, meaningful alternative text, zoom behavior, and visual accessibility.

IN THIS VIDEO

  • Lesson 14: Usability and accessibility Full course · locked
Frame from Performance, visual quality, and experience3:35
VIDEO 15 · FULL COURSE

Performance, visual quality, and experience

Investigate heavy assets, awkward layouts, and experiences that undermine customer trust.

IN THIS VIDEO

  • Lesson 15: Performance, visual quality, and experience Full course · locked
Frame from Functionality, reliability, and compatibility5:42
VIDEO 16 · FULL COURSE

Functionality, reliability, and compatibility

Check calculations, hidden retries, network problems, and responsive layouts.

IN THIS VIDEO

  • Lesson 16: Functionality, reliability, and compatibility Full course · locked
Frame from Content, localization, data, and APIs4:46
VIDEO 17 · FULL COURSE

Content, localization, data, and APIs

Compare claims, locales, stored data, and API behavior for contradictions.

IN THIS VIDEO

  • Lesson 17: Content, localization, data, and APIs Full course · locked
Frame from Review AI-generated code2:53
VIDEO 18 · FULL COURSE

Review AI-generated code

Ask for focused code review and challenge what the generating agent overlooked.

IN THIS VIDEO

  • Lesson 18: Review AI-generated code Full course · locked
Frame from Standards and compliance checks4:36
VIDEO 19 · FULL COURSE

Standards and compliance checks

Use AI to flag potential issues for expert review, not to certify legal compliance.

IN THIS VIDEO

  • Lesson 19: Standards and compliance checks Full course · locked
Frame from Reduce noise. Test state and relevance.3:22
VIDEO 20 · FULL COURSE

Reduce noise. Test state and relevance.

Bring known issues into the prompt, vary application state, and judge search relevance.

IN THIS VIDEO

  • Lesson 20: Reduce noise. Test state and relevance. Full course · locked
Frame from Recovery, specifications, and parallel checks2:59
VIDEO 21 · FULL COURSE

Recovery, specifications, and parallel checks

Review recovery paths and specifications, then combine complementary artifact checks.

IN THIS VIDEO

  • Lesson 21: Recovery, specifications, and parallel checks Full course · locked
Frame from Part One: the quality-checking loop2:30
VIDEO 22 · FULL COURSE

Part One: the quality-checking loop

Connect evidence, focus the questions, challenge findings, and make bugs actionable.

IN THIS VIDEO

  • Lesson 22: Part One: the quality-checking loop Full course · locked

Browse every lesson here. Only marked previews are publicly playable; watch the full course on Udemy →. Practice time is additional.

TAKE THE COURSE. PUT IT INTO PRACTICE.

A record of what you’ve learned.

Course completion certification is available only to people who have taken this course and passed its exam. All questions must be answered correctly; you can review and retry.

Submit your self-reported completion code to Jason for manual review. This is a course certificate—not professional accreditation or an automatically verified credential.

Open the exam · access code required →
Illustration only. No certificate has been issued.

Sample certificate · not issued

Sample certificate with placeholder Your Name and certificate ID SAMPLE
Buy How AI Tests Software by Jason Arbon on Amazon
OPTIONAL COMPANION REFERENCE

The course comes first.
The book is a reference.

Build your skills through the course’s guided lessons, demonstrations, and hands-on exercises. Keep the optional book nearby to look up concepts as you put those skills to work.

Buy the book on Amazon →
TOOLS & SERVICES

Put the learning to work.

CARBON — Engineering ConfidenceBuilt by testers.ai

Let me help you test. You take the credit.

I’ve packed my testing experience and methods into CARBON, an AI plugin that helps investigate bugs, build and run tests, and assess release confidence. Put it to work on your product, and take the credit.

Explore CARBON
IcebergQA

Find what your AI missed.

Add instant-on AI QA bandwidth and coverage, with human testing expertise. Start getting value in days, and get help learning to run it yourself.

Talk to IcebergQA
FREE COURSE PREVIEW

Read the transcript
Part One / AI Checks

Feedback / Errata

Spotted something unclear or incorrect? Send Jason a note. Feedback is kept separate from the course interest list.

How AI Tests Software — course overview

Read the video overview

A six-part course with Jason Arbon: 73 recorded lessons, approximately six hours, and 247 presentation slides. Five selected highlights show AI checks, interactive AI tests, contextual testing, tests expressed as intent, and the AI test harness. Practice with copy-and-paste prompts and the TestBucks example. The book is an optional companion reference.

Course participants who pass every exam question can submit their self-reported completion code to Jason for manual review. The certificate shown is a sample, not an issued credential. The video contains instrumental background music and no spoken narration.

Music: “Inspired” by Kevin MacLeod, music credit and CC BY 4.0 license.