AIPM-2 Exam Syllabus

Certification Overview

AIPM-2: AI Product Manager Professional is the professional practice-level certification in the APD Certified AI Product Manager pathway.

AIPM-2 validates the ability to apply AI-native product management practices to realistic product scenarios, from opportunity discovery to workflow design, product requirements, solution trade-offs, evaluation, adaptive delivery, responsible launch, monitoring, iteration, adoption, and value realization.

This certification is designed for product professionals who want to demonstrate practical AI product management capability beyond foundational understanding.

AIPM-2 is based on the APD AI-Native Product Manager Competency Model, which defines the professional capabilities required for product professionals to transform AI capabilities into trusted, usable, measurable, governable, and scalable product value.

AIPM-2 is the second level in the APD certification pathway:

• AIPM-1: AI Product Manager Foundation
• AIPM-2: AI Product Manager Professional
• AIPM-3: AI Product Manager Strategic Leadership

Version 1.1 Update

Version 1.1 strengthens the concept of AI-Native Product Agility at the professional practice level.

In AIPM-2, agility is treated as an applied AI product management capability, not as a traditional Agile or Scrum framework topic.

Candidates are expected to understand how to apply adaptive, evaluation-driven, governance-aware product delivery practices to AI product initiatives.

Version 1.1 specifically strengthens:

• iterative AI product discovery;
• experiment-driven product learning;
• MVP, prototype, proof of concept, pilot, launch, and scale-up distinctions;
• feedback-to-roadmap loops;
• evaluation-driven iteration;
• AI quality regression awareness;
• adaptive backlog and roadmap decisions;
• responsible launch readiness;
• balancing speed, risk, governance, quality, and value.

AIPM-2 does not test Scrum events, Agile roles, velocity, story points, or specific Agile frameworks. It tests whether candidates can apply adaptive AI product delivery and learning loops in realistic product scenarios.

Positioning of AIPM-2

AIPM-1 validates foundational product judgment.
AIPM-2 validates applied professional practice.
AIPM-3 validates strategic leadership.

AIPM-2 is designed to answer one core question:

Can the candidate apply AI-native product management practices to a realistic AI product scenario and make sound product decisions across value, workflow, experience, technology, knowledge, evaluation, adaptive delivery, and responsible launch?

AIPM-2 is not intended to certify enterprise-level AI product strategy, AI product portfolio leadership, organizational AI transformation leadership, or principal-level governance capability. Those capabilities belong primarily to AIPM-3.

AIPM-2 focuses on the product professional who can independently manage a defined AI product, AI feature, AI workflow, AI assistant, AI agent use case, GenAI application, enterprise AI use case, or AI-enabled product area under normal organizational supervision.

What AIPM-2 Validates

AIPM-2 validates that candidates can apply AI-native product management practices in realistic product contexts.

A successful AIPM-2 candidate should be able to:

• identify and qualify real AI product opportunities;
• determine whether AI is appropriate for a given problem;
• define users, business outcomes, value hypotheses, and success metrics;
• design human-AI workflows for defined product scenarios;
• define AI-native product requirements;
• collaborate effectively with AI, data, engineering, design, security, legal, business, delivery, and customer-facing stakeholders;
• define knowledge, retrieval, context, and data readiness requirements;
• evaluate product-level technical options and trade-offs;
• design AI product experience patterns involving trust, fallback, human review, and user control;
• create basic AI evaluation plans and quality gates;
• use evaluation findings, user feedback, monitoring signals, and risk findings to adjust roadmap, backlog, and product decisions;
• assess product-level risks related to responsible AI, safety, security, privacy, and governance;
• support MVP, pilot, launch, monitoring, iteration, adoption, and value realization;
• communicate AI product decisions clearly to cross-functional stakeholders.

AIPM-2 validates applied product capability. It does not require candidates to personally implement machine learning systems, fine-tune models, build infrastructure, or perform security engineering.

Who Should Earn AIPM-2

AIPM-2 is designed for professionals who are working on, transitioning into, or preparing to independently manage AI-enabled or AI-native product work.

It is especially suitable for:

• product managers working on AI-enabled products;
• product managers transitioning into AI product roles;
• GenAI application product managers;
• AI assistant or AI copilot product managers;
• AI agent or workflow product managers;
• enterprise AI product managers;
• internal tools or productivity AI product managers;
• AI platform associate product managers;
• technical product managers working with AI teams;
• product owners responsible for AI-enabled capabilities;
• agile product professionals supporting AI product delivery;
• consultants supporting AI product design and delivery;
• professionals seeking evidence of practical AI product capability.

AIPM-2 is most appropriate for candidates who already understand product management fundamentals and have either practical product experience or strong preparation through case practice, product artifacts, or AI product project work.

Recommended Prerequisites

AIPM-1 certification or equivalent foundational knowledge is strongly recommended before taking AIPM-2, but AIPM-1 is not strictly required.

Candidates are encouraged to have one or more of the following:

• AIPM-1 certification or equivalent foundational knowledge;
• product management experience;
• product owner or business analyst experience;
• agile delivery or product delivery experience;
• experience working with software, data, AI, or digital products;
• participation in an AI product, GenAI application, AI assistant, RAG, agent, automation, or analytics product initiative;
• completion of AI product case practice or portfolio preparation.

AIPM-2 is not recommended for candidates who have no product management exposure and no understanding of AI product fundamentals.

No machine learning engineering experience is required.

No Scrum or Agile certification is required.

What AIPM-2 Does Not Validate

AIPM-2 is a professional practice-level certification. It does not validate strategic or principal-level AI product leadership.

AIPM-2 does not certify that a candidate can independently perform:

• enterprise AI product portfolio strategy;
• organization-wide AI transformation leadership;
• enterprise AI governance system design;
• AI product platform strategy across multiple business units;
• principal-level AI architecture decision-making;
• advanced AI evaluation system design across a portfolio;
• AI product operating model redesign at enterprise scale;
• AI product community, methodology, or standard ownership at an organizational level;
• legal compliance interpretation;
• security engineering;
• model training, fine-tuning implementation, or algorithm optimization;
• production infrastructure engineering;
• Scrum Master, Agile Coach, or enterprise Agile transformation responsibilities.

These capabilities are addressed primarily at AIPM-3 or in specialized technical, legal, governance, security, agile delivery, or organizational transformation roles.

Recommended Assessment Format

AIPM-2 should assess applied professional practice. Therefore, APD Institute recommends a two-component assessment model.

Component A: Online Applied Assessment

Component B: Practical Case Assignment

APD Institute recommends that AIPM-2 include a practical case assignment or product artifact assessment.

Candidates may be asked to analyze a realistic AI product scenario and produce selected work outputs such as:

• AI product opportunity brief;
• human-AI workflow design;
• AI-native product requirements outline;
• knowledge, retrieval, and context requirements;
• AI evaluation plan;
• adaptive delivery and learning-loop plan;
• human review and fallback plan;
• responsible AI risk assessment;
• launch readiness checklist.

The practical case assignment may be graded as pass/fail or included as part of the final certification score, depending on APD Institute’s assessment policy.

If APD Institute chooses to launch AIPM-2 first as an online-only assessment, the practical case assignment should still be offered as a recommended portfolio practice and may become part of the certification in a future version.

Recommended Assessment Weighting

If APD Institute uses the two-component model, the recommended weighting is:

If APD Institute uses an online-only model, the exam should include a higher proportion of case-based, artifact-based, and adaptive-delivery judgment questions to preserve the applied nature of AIPM-2.

Cognitive Level Distribution

AIPM-2 assesses applied professional capability. It should not be dominated by recall or terminology questions.

AIPM-2 should test applied judgment. It should not test advanced strategic leadership at AIPM-3 level.

Question Type Distribution

AIPM-2 should contain very few pure recall questions.

Exam Domain Blueprint

AIPM-2 is organized into eight exam domains.

Exam Domains

Domain 1: AI Product Opportunity Qualification and Value Framing

Weight: 12%

Domain Purpose

This domain validates whether candidates can identify, qualify, and frame AI product opportunities in relation to user problems, business outcomes, feasibility, risk, adoption readiness, and measurable value.

Learning Objectives

Candidates should be able to:

1. identify realistic AI product opportunities from user, business, workflow, or operational problems;
2. distinguish user problems, business problems, technology ideas, and AI use cases;
3. evaluate whether AI is appropriate for a given problem;
4. define value hypotheses for AI product initiatives;
5. identify target users, beneficiaries, decision-makers, and affected stakeholders;
6. define measurable success metrics for AI product outcomes;
7. identify basic cost, latency, quality, risk, and adoption considerations;
8. avoid technology-first, model-first, or demo-driven product thinking;
9. design an evidence-based discovery approach for validating AI product opportunities;
10. determine when a product idea should proceed to experiment, prototype, MVP, pilot, or be stopped.

Key Topics

• AI opportunity discovery;
• problem framing;
• user value and business value;
• use case qualification;
• value hypothesis;
• measurable outcomes;
• feasibility, usability, viability, and risk;
• AI suitability assessment;
• product value vs. technical novelty;
• success metrics;
• cost and latency awareness;
• adoption feasibility;
• evidence-based discovery;
• AI-native product agility in discovery.

Candidates Should Be Able to Produce or Evaluate

• AI opportunity brief;
• problem statement;
• user and stakeholder definition;
• value hypothesis;
• AI suitability rationale;
• success metric proposal;
• use case prioritization rationale;
• discovery experiment proposal;
• decision rationale for proceed, pause, redesign, or stop.

Not in Scope for AIPM-2

• enterprise AI investment portfolio strategy;
• pricing and packaging ownership for complex commercial products;
• business-unit-level AI transformation strategy;
• advanced financial modeling;
• enterprise continuous discovery operating model design.

Domain 2: Human-AI Workflow and Scenario Design

Weight: 15%

Domain Purpose

This domain validates whether candidates can design practical human-AI workflows for defined AI product scenarios and adapt those workflows based on feedback, evaluation, and risk findings.

Learning Objectives

Candidates should be able to:

1. decompose business scenarios into tasks, decisions, inputs, outputs, actors, and constraints;
2. identify which tasks should be automated, augmented, reviewed, escalated, or kept human-led;
3. design human-AI workflows for defined scenarios;
4. specify AI system responsibilities and human responsibilities;
5. define human-in-the-loop, human-on-the-loop, and human-over-the-loop mechanisms;
6. design fallback, escalation, exception handling, and human review points;
7. define agentic workflow boundaries, tool-use permissions, approval steps, and auditability needs;
8. align workflow design with value, risk, user trust, operational feasibility, and performance metrics;
9. identify workflow assumptions that need validation through pilot, user feedback, monitoring, or evaluation;
10. adapt workflow design based on learning from real usage and evaluation findings.

Key Topics

• scenario modeling;
• task-level work analysis;
• human-AI workflow design;
• human-AI task allocation;
• decision flow;
• human oversight;
• fallback and escalation;
• agentic workflow;
• tool use and action boundaries;
• approval before action;
• exception handling;
• accountability and audit trail;
• workflow performance metrics;
• workflow validation;
• iterative workflow improvement.

Candidates Should Be Able to Produce or Evaluate

• human-AI workflow map;
• task allocation matrix;
• agent responsibility matrix;
• fallback and escalation design;
• human review rules;
• decision flow diagram;
• workflow risk checklist;
• workflow learning-loop plan.

Not in Scope for AIPM-2

• enterprise-wide operating model redesign;
• multi-business-unit workflow transformation;
• advanced multi-agent orchestration architecture;
• organizational change strategy ownership;
• detailed workflow facilitation methodology.

Domain 3: AI-Native Product Requirements and Experience Design

Weight: 14%

Domain Purpose

This domain validates whether candidates can define AI-native product requirements and design trustworthy product experiences for realistic AI product scenarios.

Learning Objectives

Candidates should be able to:

1. translate AI product opportunities and workflows into AI-native product requirements;
2. define functional requirements, AI behavior requirements, output requirements, quality requirements, and constraints;
3. specify user interaction patterns for AI assistants, copilots, agents, decision-support tools, or generative experiences;
4. define prompt, instruction, and output behavior at a product specification level;
5. design trust, transparency, user control, user correction, fallback, and human review mechanisms;
6. identify acceptance criteria for AI behavior and user experience;
7. define requirements for explainability, source citation, limitation communication, and user consent where appropriate;
8. balance usability, reliability, safety, operational cost, user trust, and delivery speed;
9. update product requirements based on evaluation findings, user feedback, production signals, and risk findings;
10. distinguish stable product requirements from assumptions that require experimentation or iterative validation.

Key Topics

• AI-native PRD;
• AI behavior specification;
• output requirements;
• acceptance criteria;
• conversation and interaction flow;
• prompt and instruction requirements;
• trustworthy product experience;
• user control and correction;
• transparency and limitation communication;
• source citation and grounding behavior;
• fallback experience;
• human review UX;
• product constraints and non-functional requirements;
• requirement iteration;
• assumption validation.

Candidates Should Be Able to Produce or Evaluate

• AI-native PRD excerpt;
• AI behavior specification;
• prompt and output requirement;
• trust and transparency design;
• fallback experience specification;
• human review requirement;
• acceptance criteria for AI output;
• requirement update based on evaluation results.

Not in Scope for AIPM-2

• advanced UX research program design;
• detailed visual design specification;
• full end-to-end enterprise product requirements ownership;
• advanced AI product design system governance.

Domain 4: AI Solution Architecture and Technical Trade-Offs for Product Managers

Weight: 13%

Domain Purpose

This domain validates whether candidates can collaborate with technical teams and make informed product decisions about AI solution approaches, technical trade-offs, feasibility, and delivery implications.

Learning Objectives

Candidates should be able to:

1. compare common AI solution approaches such as prompting, RAG, fine-tuning, tool calling, agent workflows, rules-based logic, and hybrid systems;
2. identify when RAG, tool use, agentic workflow, or rules-based automation may be appropriate;
3. evaluate product-level trade-offs across quality, cost, latency, reliability, scalability, maintainability, security, and user experience;
4. recognize risks related to model choice, vendor dependency, data access, integrations, permissions, and operational complexity;
5. define product-level architecture questions for technical teams;
6. distinguish product decisions from engineering implementation decisions;
7. participate in build-buy-partner discussions;
8. communicate technical trade-offs in product and business language;
9. understand how technical changes may introduce quality regression, cost changes, latency changes, or new risks;
10. incorporate technical learning into product roadmap and iteration decisions.

Key Topics

• AI solution options;
• prompting vs. RAG vs. fine-tuning;
• tool calling;
• agentic workflows;
• rules-based and hybrid approaches;
• build vs. buy vs. partner;
• model selection considerations;
• vendor risk and lock-in;
• latency and cost trade-offs;
• reliability and scalability;
• security exposure;
• integration complexity;
• product architecture decision-making;
• technical change impact;
• roadmap adaptation.

Candidates Should Be Able to Produce or Evaluate

• AI solution concept;
• technical feasibility questions;
• product architecture brief;
• build-buy-partner analysis;
• AI cost and latency trade-off summary;
• integration and dependency map;
• technical learning summary for roadmap decisions.

Not in Scope for AIPM-2

• model training;
• neural network architecture;
• fine-tuning implementation;
• vector database configuration;
• infrastructure deployment;
• security engineering;
• algorithm research;
• detailed cloud architecture design.

Domain 5: Enterprise Knowledge, Context, Data, and Retrieval Requirements

Weight: 10%

Domain Purpose

This domain validates whether candidates can define and evaluate the data, knowledge, context, retrieval, and governance requirements needed for AI products to perform reliably in enterprise contexts.

Learning Objectives

Candidates should be able to:

1. identify the data, knowledge, documents, user context, operational signals, and business rules needed by an AI product;
2. assess knowledge source quality, coverage, freshness, sensitivity, permissions, and reliability;
3. define knowledge base and retrieval requirements;
4. define requirements for grounding, citations, source traceability, and permission-aware access;
5. recognize risks from outdated, incomplete, conflicting, unauthorized, or untrusted knowledge sources;
6. define feedback loops for improving knowledge and retrieval quality;
7. collaborate with data, engineering, security, legal, subject matter experts, and operations teams on data and knowledge readiness;
8. distinguish data readiness, knowledge readiness, and context readiness;
9. identify when knowledge, data, or retrieval changes require re-evaluation;
10. incorporate retrieval and knowledge-quality learning into product iteration.

Key Topics

• data readiness;
• knowledge readiness;
• context engineering;
• enterprise knowledge assets;
• RAG requirements;
• retrieval quality;
• source reliability and freshness;
• access control;
• permission-aware retrieval;
• data sensitivity;
• metadata and source traceability;
• grounding and citation;
• knowledge update workflows;
• feedback loops;
• retrieval improvement;
• knowledge-change regression risk.

Candidates Should Be Able to Produce or Evaluate

• data and knowledge requirements;
• knowledge readiness checklist;
• context requirements;
• retrieval quality requirements;
• source reliability assessment;
• permission-aware access requirements;
• knowledge feedback loop;
• retrieval improvement plan.

Not in Scope for AIPM-2

• detailed data pipeline engineering;
• embedding model selection;
• retrieval ranking algorithm design;
• enterprise knowledge architecture ownership;
• database administration;
• data engineering implementation.

Domain 6: AI Evaluation, Experimentation, and Product Intelligence Planning

Weight: 16%

Domain Purpose

This domain validates whether candidates can design practical evaluation and experimentation plans for AI products and use evaluation insights to guide product decisions, learning loops, and iteration.

Learning Objectives

Candidates should be able to:

1. define AI product evaluation objectives;
2. distinguish model evaluation, product evaluation, user experience evaluation, business value evaluation, and risk evaluation;
3. select appropriate evaluation metrics for defined AI product scenarios;
4. design basic evaluation plans using test sets, golden datasets, human review rubrics, acceptance thresholds, and quality gates;
5. define evaluation requirements for hallucination, groundedness, relevance, task success, safety, latency, cost, and user satisfaction;
6. design pilot evaluation and phased rollout learning plans;
7. identify regression risks after changes to models, prompts, retrieval sources, tools, workflows, or knowledge bases;
8. use evaluation findings to inform roadmap, backlog, UX, prompt, retrieval, model, knowledge, workflow, and governance decisions;
9. define learning loops that connect evaluation results, user feedback, monitoring signals, and product decisions;
10. distinguish experimentation for discovery, evaluation for readiness, and monitoring for continuous improvement.

Key Topics

• AI product evaluation;
• evaluation plan;
• experimentation;
• golden dataset;
• test set design;
• human review rubric;
• acceptance threshold;
• quality gate;
• offline and online evaluation;
• pilot evaluation;
• task success;
• groundedness;
• hallucination rate;
• relevance;
• safety metrics;
• latency and cost metrics;
• user satisfaction;
• business value metrics;
• regression testing;
• product intelligence;
• evaluation-to-roadmap loop;
• learning loop;
• experiment-driven development.

Candidates Should Be Able to Produce or Evaluate

• AI evaluation plan;
• metrics framework;
• human review rubric;
• quality gate checklist;
• pilot learning plan;
• experiment design;
• evaluation dashboard concept;
• iteration recommendation based on evaluation results;
• feedback-to-roadmap loop.

Not in Scope for AIPM-2

• advanced statistical experiment design;
• automated evaluation system engineering;
• enterprise-wide evaluation platform ownership;
• red-team program leadership;
• model benchmarking research;
• production observability architecture design.

Domain 7: Responsible AI, Safety, Security, and Product Risk Management

Weight: 12%

Domain Purpose

This domain validates whether candidates can identify, assess, mitigate, and document product-level AI risks for realistic AI product scenarios, while supporting responsible product learning and delivery.

Learning Objectives

Candidates should be able to:

1. perform product-level AI risk identification for defined use cases;
2. classify basic risk factors related to users, decisions, data, safety, privacy, fairness, security, and operational impact;
3. define human oversight, escalation, appeal, override, and accountability mechanisms;
4. identify LLM and agent-related risks such as prompt injection, sensitive information disclosure, unsafe tool use, excessive agency, insecure output handling, and unauthorized access;
5. define transparency, disclosure, limitation communication, and contestability requirements where appropriate;
6. collaborate with security, privacy, legal, compliance, risk, data, engineering, and business stakeholders;
7. document risk decisions, mitigations, assumptions, and launch readiness concerns;
8. apply responsible AI principles without treating governance as a late-stage compliance checklist;
9. define governance checkpoints that support safe iteration rather than blocking learning;
10. identify when new product learning, workflow changes, or technical changes require renewed risk review.

Key Topics

• responsible AI;
• product-level AI risk assessment;
• risk classification;
• human oversight;
• privacy and sensitive data;
• fairness and impact;
• transparency and disclosure;
• explainability and contestability;
• prompt injection;
• unsafe tool use;
• excessive agency;
• insecure output handling;
• access control;
• auditability and traceability;
• incident awareness;
• governance review;
• responsible launch readiness;
• governance-aware iteration;
• adaptive risk review.

Candidates Should Be Able to Produce or Evaluate

• AI product risk assessment;
• responsible AI checklist;
• human oversight design;
• transparency and disclosure plan;
• LLM application security risk checklist;
• product-level mitigation plan;
• launch governance checklist;
• risk review trigger list.

Not in Scope for AIPM-2

• detailed legal compliance analysis;
• jurisdiction-specific regulatory interpretation;
• enterprise AI governance program leadership;
• security engineering;
• penetration testing;
• advanced threat modeling;
• formal AI audit leadership.

Domain 8: Adaptive AI Product Delivery, Launch Readiness, Adoption, and Value Realization

Weight: 8%

Domain Purpose

This domain validates whether candidates can support adaptive delivery, launch, monitoring, adoption, and value realization of defined AI product initiatives.

Learning Objectives

Candidates should be able to:

1. define MVP and pilot scope for an AI product initiative;
2. distinguish prototype, proof of concept, experiment, MVP, pilot, production launch, scale-up, and continuous improvement;
3. apply iterative discovery and delivery practices to AI product initiatives;
4. use evaluation results, user feedback, monitoring signals, and risk findings to adjust product roadmap and backlog;
5. recognize why AI product teams require shorter learning cycles due to uncertainty in model behavior, data quality, user trust, risk, and business value;
6. balance delivery speed with evaluation, governance, safety, and production readiness;
7. identify launch readiness requirements for AI products;
8. define monitoring requirements for quality, safety, latency, cost, usage, feedback, and incidents;
9. plan basic adoption and enablement activities;
10. define business value tracking after launch;
11. support iteration based on monitoring, evaluation, user feedback, and business outcomes.

Key Topics

• adaptive AI product delivery;
• AI-native product agility;
• iterative discovery;
• experiment-driven product development;
• prototype vs. PoC vs. experiment vs. MVP vs. pilot vs. production;
• launch readiness;
• rollout planning;
• monitoring and observability requirements;
• incident awareness;
• feedback loop;
• learning loop;
• product backlog adaptation;
• roadmap adaptation;
• AI quality regression;
• continuous validation;
• governance-aware agility;
• user adoption;
• enablement;
• customer success;
• value realization;
• iteration planning.

Candidates Should Be Able to Produce or Evaluate

• MVP scope;
• pilot plan;
• experiment plan;
• launch readiness checklist;
• monitoring requirements;
• adoption plan;
• feedback loop design;
• roadmap adaptation recommendation;
• value realization report outline;
• post-launch iteration plan.

Not in Scope for AIPM-2

• enterprise-wide adoption strategy;
• full customer success operating model design;
• product portfolio lifecycle management;
• organizational transformation leadership;
• platform operating model ownership;
• Scrum Master or Agile Coach practice assessment;
• enterprise Agile transformation design.

Cross-Cutting Themes

AIPM-2 includes several cross-cutting themes that may appear across multiple exam domains.

Applied Product Judgment

Candidates must demonstrate that they can make product decisions in realistic AI product scenarios, not merely recall definitions.

Value Before Technology

AI product work should begin with user problems, business outcomes, workflow needs, and measurable value, not with the newest model or tool.

AI-Native Product Agility

AI-native product management requires adaptive learning, iterative delivery, continuous evaluation, and responsible scaling.

Candidates should understand how to use product discovery, experiments, pilots, evaluation, monitoring, user feedback, and risk findings to improve AI product value, quality, safety, trust, and readiness.

In AIPM-2, agility is assessed as applied product practice, not as traditional Agile framework knowledge.

Human-AI Workflow as the Core Unit of Design

Candidates must be able to reason about how humans, AI systems, agents, data, tools, and enterprise workflows interact.

Evaluation-Driven Product Development

Candidates must understand that AI product quality requires systematic evaluation, quality gates, regression awareness, monitoring, and iteration.

Responsible Launch

Candidates must integrate responsible AI, safety, security, privacy, human oversight, and governance into product decisions before launch.

Enterprise Context

Candidates must recognize that enterprise AI products require data readiness, knowledge readiness, access control, stakeholder alignment, operational readiness, and value measurement.

Cross-Functional Collaboration

Candidates must understand when to involve engineering, AI, data, design, security, privacy, legal, compliance, business, operations, customer success, delivery, and governance stakeholders.

Mapping to the APD AI-Native Product Manager Competency Model

AIPM-2 focuses on applied professional practice across the APD AI-Native Product Manager Competency Model.

AIPM-2 is deeper than AIPM-1 because it requires candidates to apply concepts to realistic product scenarios. It is narrower than AIPM-3 because it does not require strategic leadership across portfolios, platforms, organizations, or enterprise governance systems.

Relationship to AI-Native Organizational Capability

AIPM-2 aligns with APD Institute’s broader view of AI-native organizational capability.

AI-native organizations require more than individual AI tool usage. They require redesigned workflows, accountable operating models, enterprise data and knowledge infrastructure, trustworthy governance, workforce capability, operational discipline, adaptive learning, and measurable value creation.

AIPM-2 prepares product professionals to contribute to this organizational capability by applying AI-native product management practices at the product, feature, workflow, or defined use-case level.

The AIPM-2 candidate should be able to support:

• value-based AI use case qualification;
• task-level and workflow-level product design;
• human-AI collaboration patterns;
• adaptive AI product delivery;
• experiment and evaluation-driven iteration;
• AI product lifecycle planning;
• data, knowledge, and retrieval readiness;
• evaluation and product readiness;
• product-level risk management;
• pilot, launch, monitoring, adoption, and value realization.

Recommended Passing Candidate Standard

A passing AIPM-2 candidate should demonstrate that they can:

1. identify and qualify realistic AI product opportunities;
2. define user problems, business outcomes, value hypotheses, and success metrics;
3. determine whether AI is appropriate for a defined product problem;
4. design practical human-AI workflows for defined scenarios;
5. define AI-native product requirements and experience requirements;
6. evaluate product-level technical options and trade-offs;
7. define knowledge, context, data, and retrieval requirements;
8. design practical AI evaluation plans and quality gates;
9. use evaluation, feedback, monitoring, and risk findings to support product iteration;
10. identify product-level risks and responsible AI requirements;
11. define launch readiness and monitoring requirements;
12. support adaptive delivery, MVP, pilot, launch, feedback, iteration, adoption, and value realization;
13. communicate product decisions clearly to cross-functional stakeholders.

A passing candidate should be capable of independently managing a defined AI product area or use case under normal organizational supervision.

Recommended Preparation

Candidates preparing for AIPM-2 should focus on:

1. reviewing the APD AI-Native Product Manager Competency Model;
2. understanding the AIPM-1 foundation-level concepts;
3. practicing AI product opportunity qualification;
4. writing AI product opportunity briefs;
5. designing human-AI workflows;
6. drafting AI-native product requirements;
7. defining knowledge, retrieval, and context requirements;
8. creating basic AI evaluation plans;
9. designing product learning loops;
10. identifying AI product risks and responsible launch requirements;
11. analyzing realistic AI product cases;
12. preparing product artifacts such as PRD excerpts, workflow maps, evaluation plans, delivery plans, and risk assessments.

Candidates should practice applying concepts to realistic product scenarios rather than memorizing terminology.

Candidates should not prepare by memorizing Agile framework details. AIPM-2 focuses on adaptive AI product delivery and evaluation-driven product learning.

Recommended AIPM-2 Practical Case Assignment

The following practical case assignment format is recommended for AIPM-2.

Case Scenario

Candidates are given a realistic AI product scenario, such as:

• customer support AI assistant;
• enterprise knowledge copilot;
• HR policy assistant;
• sales enablement copilot;
• AI agent for service ticket triage;
• document review assistant;
• developer support assistant;
• internal productivity AI tool;
• AI-powered analytics assistant.

Candidate Deliverables

Candidates may be asked to submit selected artifacts:

1. AI product opportunity brief;
2. target user and problem statement;
3. value hypothesis and success metrics;
4. human-AI workflow map;
5. AI-native product requirements excerpt;
6. knowledge, data, context, and retrieval requirements;
7. AI evaluation plan;
8. learning-loop and iteration plan;
9. human review, fallback, and escalation plan;
10. responsible AI risk assessment;
11. launch readiness and monitoring plan.

Evaluation Criteria

The practical case should be evaluated against:

• problem clarity;
• AI suitability;
• value hypothesis quality;
• workflow design quality;
• human oversight design;
• product requirement clarity;
• knowledge and retrieval readiness;
• evaluation plan quality;
• learning-loop quality;
• adaptive delivery thinking;
• responsible AI and risk awareness;
• launch readiness;
• cross-functional collaboration awareness;
• clarity of communication.

AIPM-2 practical assessment should reward sound product judgment, not excessive documentation volume.

Suggested Portfolio Evidence

AIPM-2 candidates are encouraged to build portfolio evidence that demonstrates applied AI product capability.

Recommended portfolio artifacts include:

• AI product opportunity brief;
• AI use case qualification;
• human-AI workflow design;
• AI-native PRD excerpt;
• prompt and output requirements;
• knowledge and retrieval requirements;
• AI evaluation plan;
• human review rubric;
• responsible AI risk assessment;
• experiment or pilot plan;
• learning-loop and feedback-to-roadmap plan;
• pilot or launch readiness checklist;
• value realization report.

These artifacts may support hiring, promotion, internal mobility, or future AIPM-3 preparation.

Employer Interpretation

For employers, AIPM-2 indicates that a candidate has applied AI product management capability and can work on realistic AI product scenarios beyond foundational awareness.

AIPM-2 suggests that a candidate can:

• qualify AI product opportunities;
• design defined human-AI workflows;
• define AI-native requirements;
• collaborate with technical and non-technical stakeholders;
• define evaluation and launch readiness requirements;
• apply adaptive AI product delivery practices;
• use learning loops to support iteration;
• identify product-level AI risks;
• support MVP, pilot, launch, monitoring, adoption, and value realization.

AIPM-2 should be considered a meaningful professional signal for defined AI product roles, but it should still be combined with interviews, portfolio review, work experience, and role-specific assessment.

Suggested Role Fit

Relationship to AIPM-1 and AIPM-3

AIPM-1: AI Product Manager Foundation

AIPM-1 validates foundational product judgment required to understand, evaluate, and participate in AI-native product work under guidance.

AIPM-1 focuses on foundational concepts, terminology, basic scenario reasoning, AI-native product agility awareness, product readiness awareness, and responsible AI awareness.

AIPM-2: AI Product Manager Professional

AIPM-2 validates applied AI product management capability in realistic product scenarios.

AIPM-2 focuses on opportunity qualification, human-AI workflow design, AI-native requirements, product-level technical trade-offs, evaluation planning, adaptive delivery, responsible launch readiness, delivery, monitoring, iteration, adoption, and value realization.

AIPM-3: AI Product Manager Strategic Leadership

AIPM-3 validates strategic leadership in complex AI product environments.

AIPM-3 focuses on AI product strategy, portfolio thinking, platform strategy, operating model alignment, adaptive governance, evaluation systems, enterprise adoption, value realization, and organizational AI product capability building.

AIPM-2 prepares candidates for AIPM-3, but it does not attempt to certify AIPM-3-level strategic leadership.

Sample Question Patterns

AIPM-2 questions may follow these patterns.

Pattern 1: Applied Scenario Judgment

Tests whether candidates can apply product judgment to realistic AI product situations.

Example topics:

• whether an AI use case is worth pursuing;
• what should be clarified before starting development;
• what risk should be addressed before launch;
• what metric best reflects product value;
• when to run an experiment, pilot, or launch.

Pattern 2: Human-AI Workflow Design

Tests whether candidates can reason about task allocation, human review, escalation, agent boundaries, and workflow learning.

Example topics:

• where human approval is required;
• what actions an AI agent should or should not take;
• how exception handling should work;
• where accountability should sit;
• how workflow should adapt after evaluation findings.

Pattern 3: Product Artifact Interpretation

Tests whether candidates can evaluate product artifacts.

Example artifacts:

• AI PRD excerpt;
• workflow map;
• evaluation plan;
• learning-loop plan;
• risk checklist;
• launch readiness checklist;
• metrics dashboard.

Pattern 4: Technical Trade-Off Reasoning

Tests whether candidates can make product-level decisions involving technical options.

Example topics:

• prompting vs. RAG;
• RAG vs. fine-tuning;
• tool calling vs. workflow automation;
• cost vs. quality;
• latency vs. user experience;
• build vs. buy;
• technical regression risk.

Pattern 5: Evaluation and Launch Readiness

Tests whether candidates can assess whether an AI product is ready for pilot or production.

Example topics:

• evaluation gaps;
• quality gates;
• human review rubrics;
• monitoring requirements;
• regression risk;
• pilot learning goals;
• feedback-to-roadmap loop.

Pattern 6: Responsible AI and Product Risk

Tests whether candidates can identify risks and propose appropriate product-level mitigations.

Example topics:

• privacy;
• sensitive data;
• prompt injection;
• unsafe tool use;
• high-risk decisions;
• transparency;
• human oversight;
• auditability;
• risk review triggers.

Pattern 7: Adaptive AI Product Delivery

Tests whether candidates can apply AI-native product agility.

Example topics:

• prototype vs. PoC vs. MVP vs. pilot vs. production;
• when to iterate, pause, scale, or redesign;
• how evaluation changes roadmap decisions;
• how to balance speed and governance;
• how to adapt backlog based on feedback and monitoring.

Representative Sample Questions

The following examples illustrate the intended style and difficulty of AIPM-2 questions. They are not final exam items.

Sample Question 1: Opportunity Qualification

A company wants to build a generative AI assistant for customer support. The business goal is to reduce support cost, but the team has not identified which customer issues the assistant should handle or how quality will be measured.

What should the product manager do first?

1. Select the most advanced language model available
   B. Define target support scenarios, user value, business outcomes, risk level, and success metrics
   C. Ask engineering to build a chatbot prototype immediately
   D. Purchase a vector database and upload all support documents

Correct Answer: B

Sample Question 2: Human-AI Workflow

An AI agent is proposed to automatically approve customer refund requests. Some refunds are low value and routine, while others involve fraud risk or policy exceptions.

What is the best product design approach?

1. Allow the agent to approve all refunds to maximize automation
   B. Block the agent from helping with refunds entirely
   C. Let the agent handle routine low-risk cases within defined limits and escalate exceptions for human review
   D. Ask users to decide whether the agent should approve each case

Correct Answer: C

Sample Question 3: Evaluation Planning

A product team says the AI assistant is ready for launch because it performed well in a live demo. What is the strongest product management concern?

1. A demo is not enough evidence of quality, safety, reliability, or readiness across expected and edge cases
   B. A demo always proves the model is ready for production
   C. The product should be launched before competitors respond
   D. Evaluation is mainly the engineering team’s responsibility and does not affect product decisions

Correct Answer: A

Sample Question 4: Knowledge and Retrieval

An internal policy assistant gives outdated answers because it retrieves old policy documents that are still stored in the knowledge base.

Which product requirement is most directly missing?

1. A friendlier assistant tone
   B. A larger model
   C. Knowledge freshness, source governance, and retrieval quality requirements
   D. A longer onboarding tutorial

Correct Answer: C

Sample Question 5: Responsible Launch

An AI sales assistant can access CRM data, generate emails, and send them to customers. What should the product manager ensure before launch?

1. The assistant uses the most creative writing style possible
   B. The assistant has permission boundaries, human approval for outbound actions, audit logs, privacy controls, and evaluation criteria
   C. The assistant sends emails automatically to maximize productivity
   D. The assistant ignores CRM permissions because the user already has access

Correct Answer: B

Sample Question 6: Adaptive Delivery

A team completed a successful prototype of an AI knowledge assistant. Early users like the concept, but evaluation shows weak groundedness on policy questions and inconsistent answers when the knowledge base is updated.

What should the product manager do next?

1. Launch immediately because users liked the prototype
   B. Redesign the learning loop by improving knowledge governance, evaluation criteria, quality gates, and iteration priorities before pilot expansion
   C. Replace the product team with a larger engineering team
   D. Remove evaluation to speed up delivery

Correct Answer: B

Candidate Preparation Guidance

Candidates should prepare by practicing realistic AI product scenarios and producing product artifacts.

Recommended preparation activities include:

• study the APD AI-Native Product Manager Competency Model;
• review the AIPM-1 foundation if needed;
• analyze AI product case studies;
• practice writing opportunity briefs;
• practice designing human-AI workflows;
• practice writing AI-native product requirements;
• practice defining knowledge and retrieval requirements;
• practice creating AI evaluation plans;
• practice designing adaptive delivery and learning loops;
• practice product-level risk assessment;
• practice launch readiness analysis;
• review responsible AI and LLM application security risks;
• prepare portfolio artifacts for future career use.

Candidates should avoid focusing only on terminology memorization. AIPM-2 requires applied product judgment.

Public Certification Statement

The following statement may be used on APD Institute certification pages:

AIPM-2: AI Product Manager Professional validates applied AI product management capability in realistic product scenarios. It is designed for product professionals who need to identify and qualify AI product opportunities, design human-AI workflows, define AI-native product requirements, collaborate with technical and business stakeholders, create evaluation plans, apply adaptive AI product delivery practices, assess product-level AI risks, and support responsible launch, monitoring, iteration, adoption, and value realization.

AIPM-2 is the professional practice level in the APD Certified AI Product Manager pathway and provides the foundation for AIPM-3: AI Product Manager Strategic Leadership.

Certification Notice

AIPM-2: AI Product Manager Professional is a professional certification issued by APD Institute. It is based on the APD AI-Native Product Manager Competency Model.

This certification does not represent a government license or a formally recognized international standard. APD Institute may update the certification structure, exam blueprint, validity rules, assessment policies, practical assignment requirements, and exam content over time.

Summary

AIPM-2 is designed to validate applied professional readiness for AI product management.

It confirms that candidates can apply AI-native product management practices to realistic AI product scenarios, including opportunity qualification, human-AI workflow design, AI-native product requirements, technical trade-off reasoning, enterprise knowledge and retrieval requirements, AI evaluation planning, adaptive AI product delivery, responsible AI risk management, launch readiness, adoption, iteration, and value realization.

AIPM-2 is the professional practice level of the APD Certified AI Product Manager pathway.