The AI Product Management Handbook
A practical guide to deciding what to build, proving it works, and operating it responsibly. The handbook connects enduring product judgement with AI evaluation, economics, interaction design, agent operations, and team design.
It is written for product managers and leaders, with useful paths for designers, engineers, and founders. You do not need a machine-learning background. You do need to care about customer outcomes, evidence, and the consequences of what ships.
32 chapters · Version 3.1 · Last updated 17 July 2026
Product Principles
The mental models that shape how high-performing product teams think, operate, and make decisions.
Outcome-Driven ThinkingUpdated
How product teams define customer and business outcomes, connect bets to evidence, and avoid mistaking faster delivery for meaningful progress.
Empowerment Needs Boundaries and AccountabilityUpdated
How to give teams authority over problems and delivery while keeping risk, escalation, and outcome ownership explicit.
Customer ObsessionUpdated
How product teams combine customer conversations, observation, behavioural evidence, and AI synthesis to find problems worth solving.
Product Signal Beats Product HopeUpdated
How to separate a durable customer instinct from a fragile solution idea, test the right uncertainty, and stop weak bets before they consume the roadmap.
AI-First, Human-CentredUpdated
What AI-first means in production, when AI-enhanced products are enough, and how to calibrate architecture, trust, economics, and autonomy.
AI Cannot Carry Your AccountabilityUpdated
The individual, team, and leadership responsibilities that remain when AI produces the work or takes the action.
AI Productivity Is a Work Design ProblemNew
Why faster output creates more review, more scope, and more pressure unless leaders redesign the work around sustainable capacity.
Product Lifecycle & Process
The operational playbook from discovery through delivery, launch, and continuous optimisation.
Discovery Is an Evidence SystemUpdated
How to run a problem queue, choose the right discovery artefact, test AI-specific risks, and turn feedback into decisions.
Business Viability, Defensibility, and AI EconomicsUpdated
How to test AI product defensibility, model costs and review load, choose pricing, and prove the economics of a production workflow.
Planning When Building Is CheapUpdated
How AI product teams limit concurrent bets, plan around evidence and operating constraints, and keep strategy clear when implementation gets cheaper.
Execution Starts With the Right ArtefactUpdated
How AI-enabled teams move from evidence to production using risk-shaped readiness, explicit context, review capacity, and progressive release.
AI Go-to-Market: Distribution, Trust, and GrowthUpdated
How AI products earn distribution, prove outcomes, build buyer trust, enable technical sales, and operate agentic growth loops.
AI Product MetricsUpdated
A practical measurement system for AI product adoption, trust calibration, value delivery, escalation patterns, and cost per outcome.
AI Governance for Regulated EnvironmentsUpdated
A risk-tiered AI governance framework covering ownership, data, security, approvals, monitoring, incidents, and regulated product decisions.
AI Product Architecture & Operations
The AI-specific technical decisions that separate production AI products from prototypes.
Model Harnesses, Routing, and SubstitutabilityUpdated
How to build the context, tool, permission, memory, evaluation, and routing layer around models without creating unnecessary orchestration.
Agentic AI Product PatternsUpdated
What makes a workflow agentic, how reliability compounds across steps, and the production patterns that survive real use.
Evaluation Frameworks as Product InfrastructureUpdated
How product teams define AI quality, build representative test sets and graders, monitor drift, and connect evaluation results to release decisions.
AI UX Is the Contract Between Human and MachineUpdated
How to design AI work surfaces, shared control, uncertainty, provenance, authenticity, and recovery across different levels of autonomy.
Every Agent Needs an OwnerNew
The operating model for agent ownership, context maintenance, permissions, evaluation, review capacity, and retirement.
Production Playbooks
Focused operating guides for agent safety, security, voice, physical AI, and other distinct production constraints.
The Agentic Safety Inspection: An Operational PlaybookUpdated
Why final-answer testing is insufficient for agents, and how to inspect operational stability, budget control, and behavioural boundaries.
Agentic Security Starts With the Blast RadiusNew
A production playbook for prompt injection, least-privilege tools, sandboxing, approvals, adversarial testing, and agent incident response.
Voice Agents in ProductionUpdated
A production playbook for latency, conversation states, human handoff, quality evaluation, and voice-agent economics.
Shipping Physical AI: Hardware, Safety, and ScaleNew
How to manage physical AI products across hardware iteration, embodied safety, simulation, manufacturing, supply chains, and field operations.
Roles, Competencies & Organisation
The product builder role, competency model, team design, adoption, and career practices for the AI era.
The Product Builder Owns the Learning LoopUpdated
The product-builder role combines problem shaping, artefact fluency, evaluation, and operational ownership to move from uncertainty to evidence.
The Product Competency ModelUpdated
A product competency framework covering customer insight, strategy, execution, leadership, commercial judgement, and the AI fluency modern roles require.
AI-Native Teams Need Generalists and SpecialistsUpdated
How to design AI-enabled teams around a generalist core, specialist depth, agent stewardship, review capacity, and the risk in the work.
AI Adoption Is an Operating Model ChangeNew
How to move AI from personal experimentation into adopted workflows through enablement, incentives, operating controls, and measured value.
The AI Fluency SpectrumUpdated
A three-stage AI fluency framework for personal output, shared systems, and redesigning how work happens across a product team.
Influence Is Internal Product DiscoveryNew
How product leaders understand decision-makers, build domain authority, surface disagreement, and improve consequential decisions without resorting to politics.
Taste Is a System, Not a VibeUpdated
How product taste combines judgement, cultural context, empathy, systems thinking, and restraint, and how to develop it deliberately.
Hiring Product BuildersUpdated
A structured hiring playbook for assessing recent evidence, product judgement, AI fluency, and the ability to own a learning loop.
Plan for Scenarios, Not One AI FutureUpdated
How to build a durable product career through task analysis, scenario planning, recent evidence, adjacent depth, and quarterly review.