01, The Role
01
Uncertainty Management at Product Level
AI outputs are probabilistic. The AI PM owns acceptance criteria for systems whose correctness cannot be binary-validated, defining thresholds, fallback behaviours, and human review triggers, not just pass/fail test cases.
02
Risk Surface Ownership
Model drift, hallucination rates, adversarial input exposure, data provenance, and bias across protected attribute groups. AI risk is continuous and evolves post-deployment as data distributions shift.
03
Regulatory Compliance Translation
Converting obligations under the EU AI Act, NIST AI RMF, and ISO 42001 into product requirements and acceptance criteria. Compliance is a product function, not a downstream legal review.
04
Post-Deployment Monitoring as First-Class Work
AI systems degrade and shift. Monitoring is not operational overhead assigned to engineering. It is a product responsibility with defined escalation thresholds, review cadences, and named accountability owners.
05
Evaluation Framework Design
Defining how AI outputs are assessed pre-deployment and monitored continuously post-deployment, including eval metric selection, human-in-the-loop thresholds, and structured review of model behaviour.
06
Vendor & Model Governance
Evaluating build vs. fine-tune vs. API decisions against capability, cost, auditability, and data sovereignty requirements. Third-party model dependencies introduce supply chain risk the AI PM must document and mitigate.
Governance Note

The AI PM retains all core product management responsibilities: problem discovery, prioritisation, roadmap governance, and stakeholder alignment. These do not disappear, they become structurally more complex because the underlying system is non-deterministic and failure modes are novel. The role extends the PM's accountability surface; it does not replace foundational PM competency.

02, Learning Roadmap
I
Phase I Foundations
Weeks 1 – 6
Learning Focus
Resource
AI/ML literacy for non-engineers: model types, training pipelines, inference, limitations
Generative AI fundamentals: LLMs, embeddings, retrieval-augmented generation, prompt engineering
Design thinking applied to probabilistic systems where outputs are non-deterministic
Lean validation methods adapted for AI: building MVPs with inherently uncertain outputs
Introduction to the AI product lifecycle: discovery through post-deployment monitoring
AI tooling landscape: categories, capability boundaries, selection and governance criteria
II
Phase II Core Practice
Weeks 7 – 16
Learning Focus
Resource
AI-specific user research: synthetic data use, bias in research samples, AI-assisted synthesis
Writing acceptance criteria for AI features: evals, thresholds, human review triggers
AI risk identification and classification: technical, ethical, regulatory, reputational dimensions
Metrics design for AI products: leading indicators, drift signals, outcome measurement
Roadmap governance for AI: managing uncertainty in sprint planning and prioritisation
Stakeholder communication: translating model behaviour to executive and legal audiences
Vendor and model governance: build vs. buy vs. fine-tune decision framework
Introduction to agentic systems: what PMs must understand about multi-agent architectures
III
Phase III Advanced Governance & Regulatory Practice
Weeks 17 – 24
Learning Focus
Resource
EU AI Act: risk tiers, conformity assessment obligations, post-market surveillance requirements
NIST AI RMF: govern, map, measure, and manage functions applied to product work
ISO 42001 AI management system: integrating governance into product operations
Sector overlays: AI implications under HIPAA, PCI DSS, DORA, SOX, and FCA guidelines
AI incident management: classification taxonomy, response protocols, regulatory notification
Responsible AI programme design: policy, review boards, accountability structures
Featured Resource
YouTube · IBM Technology
IBM Technology AI Channel
One of the most credible free AI education resources available. IBM's channel covers LLMs, RAG, AI governance, model evaluation, and enterprise AI architecture, all explained without oversimplification. Recommended throughout Phases I and II.
Watch on YouTube ↗
03, Tool Directory
Discovery & User Research
ToolPrimary UseGovernance Note
UserTesting.comModerated and unmoderated usability testing at scaleVerify data residency for regulated sector deployments
MazeRapid prototype testing with quantitative usability metricsLimited audit trail; not suitable for compliance documentation
DovetailQualitative research synthesis, tagging, and repositoryStrong audit trail; preferred for regulated environments
Notably.aiFree TierAI-assisted theme extraction from user interviewsReview AI summaries before regulated-context documentation
HotjarFree TierBehavioural analytics: heatmaps, session recordingsGDPR consent required; review data retention policies
Tally.soFreeStructured surveys for quantitative user researchGDPR compliance depends on hosting configuration
Synthetic UsersFree TierAI-generated synthetic personas for early discoveryCannot substitute for real user research
MiroFree TierJourney mapping, affinity diagrams, stakeholder workshopsConfirm data handling before including PII in boards
Solution Design & Prototyping
ToolPrimary UseGovernance Note
FigmaFree TierUI/UX prototyping for AI product interfacesNo specific AI governance considerations
Lovable.devRapid AI-powered app prototyping from natural languageReview generated code for vulnerabilities before deployment
CursorAI-assisted code editor for technical prototypingData sent to underlying model; review confidentiality terms
v0.dev (Vercel)Free TierAI-generated UI component prototypingGenerated code requires security review before production
RetoolInternal tool builder with AI component supportStronger auditability; preferred for enterprise prototyping
ChatPRDFree TierAI-assisted PRD drafting from problem statementsMandatory human review before regulated feature documentation
AI Governance, Risk & Compliance
ToolPrimary UseGovernance Note
NIST AI RMFFreeRisk management: govern, map, measure, managePrimary reference framework for US-regulated environments
EU AI Act ToolFreeOfficial self-assessment for AI Act risk classificationLegally binding for EU markets; retain outputs as records
CREDO AIFree TierAI governance and risk documentation platformReview data handling before uploading sensitive model info
Weights & BiasesFree TierML experiment tracking, model versioning, eval loggingStrong auditability; preferred for teams building custom models
Arize AIFree TierModel monitoring: drift detection, performance alertsAligns with EU AI Act Art. 9 post-deployment monitoring
Model Cards ToolkitFreeStandardised AI model transparency documentationUse as base for any model documentation programme
Holistic AIAI bias auditing, compliance tracking, risk scoringSector-specific modules for high-risk AI classification
Market Intelligence & Industry Research
SourcePrimary UseGovernance Note
TechCrunchFreeBreaking news on AI, fintech, and enterprise software; early signal on market movementsSignal source only; verify claims before citing in regulated-context documents
Tech in AsiaFree TierTechnology and startup news focused on Asian markets; useful for APAC regulatory contextRegional lens; cross-reference with local regulatory sources for compliance work
World Economic ForumFreeGlobal policy and technology trend reports; AI governance frameworks and white papersInstitutional credibility; suitable for citing in governance documentation
GartnerMagic Quadrant vendor assessments, Hype Cycle for AI, market forecastsPaid access required for full reports; Magic Quadrant cited widely in procurement decisions
StatistaFree TierAggregated statistics on AI market size, adoption rates, and industry benchmarksVerify primary source behind each statistic before using in formal research
IDCEnterprise technology market forecasts and vendor share analysisPaid reports; widely cited in board-level and investor-facing documents
Google TrendsFreeSearch volume trends for AI governance, product security, and regulatory termsDirectional signal only; not suitable as a primary research source
McKinsey AI InsightsFreeAnnual State of AI report; enterprise AI adoption benchmarks and governance findingsInfluential with senior stakeholders; methodology notes vary by report
Accenture AI ResearchFreeIndustry-specific AI adoption reports; financial services AI governance perspectivesConsulting firm output; treat as practitioner signal rather than independent research
KPMG AI ReportsFreeAI trust, risk, and governance reports; regulated industry AI adoption benchmarksConsulting firm output; useful for board-level framing of AI governance obligations
Singularity HubFreeEmerging technology trends and long-range forecasts; AI capability trajectoriesSpeculative framing common; useful for horizon scanning, not near-term compliance planning