Product Engineering with AI cover

“AI accelerates execution; thinking still determines direction.”

Addy Osmani, Director, Google Cloud AI.
Buy now

Make Products with AI.

"Product Engineering with AI" is a guide to building AI-powered products in the modern era.

"Product Engineering with AI" is a practical guide for building AI-powered products in the modern era. It's part of a complete bundle that also includes AI in Practice (a 30-page case study collection highlighting real-world examples of AI in production across multiple industries), Building Products in the Age of AI (a 30-minute talk by Addy Osmani on designing human-centered AI experiences), and Building Products with AI (a 45-minute discussion between Addy & Hassan on building an AI-powered product from scratch).

AI is fundamentally changing how we build products. From code generation to automated testing, from intelligent UX personalization to autonomous agents, the landscape of product development is evolving rapidly. The challenge is knowing how to navigate this transformation with clarity and confidence.

You'll learn how to:

What's in the bundle

Book • 200 pages

The Book Product Engineering with AI

A comprehensive guide to building AI-powered products, from foundational concepts to real-world implementation strategies and best practices

Foundations Tools Implementation
Explore →
Book • 30 pages

The Study AI in Practice

Real-world success stories from industry leaders implementing production AI solutions

Success Stories Real World
Explore →
Screencast • 45 minutes

The Workshop Building Products with AI

Shows how Addy & Hassan built an AI-powered product (Zanuh) from scratch, teaching how to clarify your vision, choose the right tools, and break AI tasks into small chunks

Demo Conversation
Explore →
Book • Field Guide

The Field Guide Product Engineering Insights in 2026

A concise guide to how AI has reshaped product engineering by 2026—from roles and team shape to tools and workflows.

Practical Patterns AI Stack
Explore →
Book • Guide

The Claude Code Playbook Product Engineering with Claude Code

A comprehensive guide to using Claude Code as a product management superpower from installation basics to advanced workflows like parallel agents, custom commands, and MCP integrations

Workflows Use Cases Power Users
Explore →
Screencast • 30 minutes

The Talk Building Products in the Age of AI

Expert insights on designing human-centered AI experiences that users trust

Design Patterns User Trust Best Practices
Explore →

The Book

A 200-page guide to building AI-powered products, from foundational concepts to real-world implementation.

Product Engineering with AI is for anyone involved in creating digital products who wants to understand how AI is reshaping the way we work. Whether you write code, design interfaces, or manage products, you’ll find practical insights on developing an AI-first mindset for the future.

Table of Contents

Part I: Foundations

  • Chapter 1 The evolution of software engineering
  • Chapter 2 Product engineering in the age of AI
  • Chapter 3 How AI is reshaping the software engineering role
  • Chapter 4 Future skills and the road ahead

Part II: Core AI tools & technologies

  • Chapter 5 AI-powered development platforms
  • Chapter 6 AI code editors
  • Chapter 7 AI agents and autonomy
  • Chapter 8 AI APIs as product building blocks

Part III: Implementation & practice

  • Chapter 9 AI-driven team collaboration and workflows
  • Chapter 10 Prompt engineering
  • Chapter 11 AI-assisted debugging and code quality
  • Chapter 12 AI-driven UX/UI design and optimization

Part IV: Strategy & future

  • Chapter 13 AI-first product development strategies
  • Chapter 14 The future of software engineering
Chapter cover preview
Book preview - page 1
Book preview - page 2
Book preview - page 3
Book preview - page 4
Book preview - page 5
Book preview - page 6
Book preview - page 7
Book preview - page 8

The Field Guide

A concise field guide to how AI has reshaped product engineering by 2026.

Product Engineering Insights in 2026 shows how small teams use copilots and agents to move faster while keeping judgment quality and user focus high. Written for builders who want practical patterns not hype and who expect AI to be part of the core stack not an add-on.

Product Engineering Insights in 2026 cover
Product Engineering Insights preview - page 1
Product Engineering Insights preview - page 3
Product Engineering Insights preview - page 4
Product Engineering Insights preview - page 5
Product Engineering Insights preview - page 6

The Playbook

A comprehensive guide to using Claude Code as a product management superpower.

Product Engineering with Claude Code covers everything from installation basics to advanced workflows like parallel agents, custom commands, and MCP integrations. Includes practical use cases including user research synthesis, PRD drafting, competitive analysis, and data reporting, with real-world tips from power users. Whether you're a PM new to AI tools or looking to level up, this guide shows you how to turn Claude Code into an indispensable collaborator.

Product Engineering with Claude Code cover
Claude Code guide preview - page 1
Claude Code guide preview - page 2
Claude Code guide preview - page 3
Claude Code guide preview - page 4
Claude Code guide preview - page 5

The Study

A 30-page companion featuring real-world success stories from industry.

AI in Practice explores how organizations in tech and beyond are putting AI to work today. From hospitality and retail to manufacturing and creative services, these case studies highlight production systems, not just prototypes, that improve core operations.

Success stories covered

  • Wendy’s FreshAI
  • Adobe Creative Cloud & Firefly
  • Toyota’s Factory AI
  • DoorDash’s Profile Generation System
  • Spotify’s AI DJ
AI in Practice cover preview
Case study 1 - page 1
Case study 1 - page 2
Case study 2 - page 1
Case study 2 - page 2

The Workshop

A 45-minute discussion between Addy Osmani and Hassan Djirdeh.

Building Products With AI shows how we built an AI powered product (Zanuh) from scratch. Learn to always clarify your vision first, choose the right tools for the job, and break AI-powered tasks into small chunks.

The workshop is broken down into 2 focused videos (with Introduction):

Introduction Vibe coding a business Vibe coding as a team
Building Products With AI — workshop cover
Building Products With AI — Addy Osmani & Hassan Djirdeh
1:00:00

The Talk

A 30‑minute presentation by Addy Osmani.

Building Products in the Age of AI is a talk on what to avoid, how to stand out, and how to design human‑centered AI experiences people trust.

Building Products in the Age of AI — talk cover
Building Products in the Age of AI — Addy Osmani
30:00

The Authors

Heyo! We're Addy & Hassan — Engineers & Educators.

Addy is the author of a number of books including Beyond Vibe Coding and his popular Substack newsletter Elevate (over 27,000 subscribers). Hassan most recently worked with Addy to author Building Large Scale Web Apps (> 3000 copies sold).

Addy Osmani

Addy Osmani

Addy is a Director in Google Cloud AI focused on helping developers and businesses succeed with Google's AI platform. Previously, he spent nearly 14 years building developer experiences in Chrome.

Hassan Djirdeh

Hassan is a senior software engineer who has built large production web applications at organizations like Doordash, Instacart, and Shopify.

Hassan Djirdeh

The Pricing

The complete bundle for building AI-powered products. No subscriptions. One payment.

Lifetime access

Written materials delivered as ePub/PDF, videos as private YouTube links.

What’s included

  • The 200‑page book
    Product Engineering with AI
    ePub/PDF
  • The 30‑page case study
    AI in Practice
    PDF
  • The Field Guide
    Product Engineering Insights in 2026
    PDF
  • The Claude Code playbook
    Product Engineering with Claude Code
    PDF
  • 30‑minute talk by Addy Osmani
    Building Products in the Age of AI
    1 private YouTube link
  • 45-minute discussion with Addy & Hassan
    Building Products with AI
    3 private YouTube links

Pay once, own everything forever

$29.99 USD

Buy now

This bundle will be available on Leanpub

Agent Code Examples

5 production-ready ADK agent projects for Product Managers and Product Engineers

This repository contains five complete Google Agent Development Kit (ADK) projects that demonstrate real-world use cases for AI agents in product development. Each agent is fully functional, well-documented, and can be run immediately with just an API key. It's part of Product Engineering with AI.

🤖 The Five Agents

1.

PRD Studio - Product Requirements Document Generator

Location: python/agents/prd_studio/

Transform fuzzy product ideas into comprehensive, execution-ready documentation.

  • What it does: Takes a brief product idea and generates a full PRD with goals, personas, MVP scope, success metrics, backlog, and risk analysis
  • Key features: Sequential agent pipeline with specialized sub-agents, built-in PRD validator
  • Use cases: Early product planning, requirements documentation, stakeholder alignment
  • Architecture: SequentialAgent with 6 specialized sub-agents
# Try it:
cd python/agents/prd_studio
adk run .
# Prompt: "Turn this into a PRD: A Chrome extension that summarizes articles using AI for busy professionals"
2.

Experiment Copilot - A/B Test Design & Analysis

Location: python/agents/experiment_copilot/

Design experiments with power analysis, then analyze results with statistical rigor.

  • What it does: Creates experiment plans with sample size calculations, analyzes results from CSV, recommends ship/iterate/stop
  • Key features: Demonstrates ADK's AgentTool pattern, includes eval suite with 5 test scenarios
  • Use cases: A/B testing, feature rollouts, growth experiments
  • Architecture: Root coordinator with specialist agents (Stats Code Agent, Narrative Decision Agent)
# Try it:
cd python/agents/experiment_copilot
adk run .
# Prompt: "Design an A/B test to improve checkout conversion by 5%"
3.

Release Radar - Dependency Upgrade Analysis

Location: python/agents/release_radar/

Scan dependencies, assess upgrade risk, and generate migration plans.

  • What it does: Parses requirements.txt/package.json, identifies upgrade risks, creates migration checklists
  • Key features: OpenAPIToolset for GitHub integration, Tool Confirmation for safety, Reflect-and-Retry plugin
  • Use cases: Dependency management, security updates, migration planning
  • Architecture: Single LLM agent with dependency parsers and optional GitHub API tools
# Try it:
cd python/agents/release_radar
adk run .
# Prompt: "Scan data/sample_requirements.txt and tell me what upgrades are high risk"
4.

VoC Insights - Customer Feedback Analysis

Location: python/agents/voc_insights/

Transform customer feedback into prioritized roadmap recommendations.

  • What it does: Ingests feedback CSV, clusters into themes, quantifies impact, generates backlog items
  • Key features: MemoryService for trend tracking, deterministic clustering, eval suite
  • Use cases: Product discovery, roadmap prioritization, user research synthesis
  • Architecture: Single LLM agent with clustering tools and memory for cross-session analysis
# Try it:
cd python/agents/voc_insights
adk run .
# Prompt: "Analyze data/sample_feedback.csv and show me the top themes with severity ratings"
5.

Meeting Ops - Meeting Analysis & Action Tracking

Location: python/agents/meeting_ops/

Turn meeting transcripts into summaries, decision logs, and action items.

  • What it does: Parses transcripts, extracts decisions/actions, drafts follow-ups, updates trackers
  • Key features: Google API toolsets (Docs, Sheets, Gmail), Tool Confirmation, cross-meeting memory
  • Use cases: Meeting documentation, action tracking, decision logging
  • Architecture: Single LLM agent with transcript parsing and optional Google Workspace integration
# Try it:
cd python/agents/meeting_ops
adk run .
# Prompt: "Load data/sample_transcript.txt and extract decisions, action items, and draft a follow-up email"

Explore all the agents and get started with the complete repository:

View on GitHub

Recommended Resources

Essential MCPs and Skills to supercharge your AI workflow.

The Core Analogy

Think of an AI agent like a professional worker.

To do a job well, they need two things: tools to do the work and the knowledge of how to use them.

The Quick Summary

  • MCP (Model Context Protocol): These are the tools (e.g., a hammer, a calculator, or access to your GitHub and database).
  • Skills: This is the instruction manual or expertise (e.g., how to build a table or how to follow your company's specific coding style).

MCPs: The "Hardware" Connections

MCP is a standard way for an AI (like Claude Code) to plug into external systems.

  • What it does: It lets the AI "touch" things outside of its chat window, like your local files, Google Drive, or Slack.
  • The downside: If you connect 20 MCP tools, the AI gets overwhelmed because it has to keep all those tool descriptions in its "short-term memory" (context window) at all times, even if it doesn't need them yet.

Skills: The "Software" Instructions

Skills are markdown files that teach the AI specific workflows or domain knowledge.

  • What it does: Instead of telling the AI every time, "Always use camelCase and run tests after every change," you save that as a "Skill."
  • The benefit: Skills are lazy-loaded. The AI only "reads" the full manual when it realizes it needs to perform that specific task, saving its context window for the actual work.

How They Work Together

Imagine you want the AI to manage your company's database:

  1. MCP provides the connection to the database so the AI can actually send queries.
  2. Skill provides the rules (e.g., "Never delete data on Fridays" or "Always check for duplicates before inserting").
Comparison table showing the differences between MCP (Model Context Protocol) tools and Skills
Feature MCP (Tools) Skills (Knowledge)
Primary Purpose Connecting to external data/APIs Encoding workflows & expertise
Storage Usually a server or JSON config Local .md files in a .skills folder
Context Usage Always "on" and taking up memory Loaded only when needed
Analogy The phone line The person knowing what to say on the call

The Recommended MCPs and Skills

1.

Rube MCP Connector

Repo: ComposioHQ/rube

What it is: A single MCP server that acts as a gateway/proxy to 500+ other integrations (Slack, GitHub, Notion, etc.) via Composio, saving you from managing individual auth tokens for every service.

2.

Superpowers

Repo: obra/superpowers

What it is: A collection of high-leverage developer skills. This is where the /brainstorm, /plan, and /finish workflows live. It basically forces Claude to adhere to a strict software development lifecycle.

The "Official" Anthropic Skills

The following are all located in the anthropics/skills monorepo. They aren't separate repositories but specific folders within this main repo.

3.

Document Suite

Location: anthropics/skills/tree/main/skills (Look for docx, xlsx, pdf, pptx)

Note: These are the reference implementations that power Claude's official file editing capabilities.

4.

Theme Factory

Location: anthropics/skills/tree/main/skills/theme-factory

Note: Allows you to define a design system (colors, fonts, spacing) that Claude applies to all generated HTML/React artifacts.

5.

Browser Automation

Playwright Browser Automation - Model-invoked Playwright automation for testing and validating web applications. By @lackeyjb

6.

Lead Research

Lead Research Assistant - Identifies and qualifies high-quality leads by analyzing your product, searching for target companies, and providing actionable outreach strategies.

7.

Webapp Testing

Location: anthropics/skills/tree/main/skills/webapp-testing

Note: Contains the logic to drive Playwright for end-to-end testing. It teaches Claude how to write and execute these tests reliably.

8.

MCP Builder

Location: anthropics/skills/tree/main/skills/mcp-builder

Note: A meta-tool. It's a skill that teaches Claude how to write the code for new MCP servers, effectively allowing it to build its own tools.

9.

Content Research

Content Research Writer - Assists in writing high-quality content by conducting research, adding citations, improving hooks, and providing section-by-section feedback.

10.

NotebookLM

NotebookLM Integration - Lets Claude Code chat directly with NotebookLM for source-grounded answers based exclusively on uploaded documents. By @PleasePrompto

The FAQ

Answers to common questions.

Do I need to be a software engineer?

No. Product Engineering with AI is designed for builders across roles; no coding required.

How long is the main book?

The book has 14 chapters totaling just about 200 pages. This will grow as we add more content and new chapters over time.

Is the book complete?

Yes! The book is complete, though we will continue to update it and add new content over time.

What format is the book and case study?

The book is available in ePub/PDF while the case study is only available as PDF.

How will I get access to the talk, discussion, case study, etc. after purchasing the book?

Once your purchase through Leanpub is complete, you'll see a Thank You post-purchase screen and receive a Thank You email with private YouTube links to our webinar series and download links for the written materials.

How does this differ from your other book, Beyond Vibe Coding?

This book focuses on shipping products and the evolving role of product engineers building them (solo founders, team members, PMs). Beyond Vibe Coding is more about AI in the software development lifecycle.

Can I read a sample for free?

Absolutely. to get sample chapters in your inbox.

Have questions before or after your purchase?

Feel free to reach out to us anytime.

Copyright © 2026 Addy & Hassan