AIPM — APMG AI-Driven Project Manager Quick Reference

Last revised: July 1, 2026

Compact exam-prep reference for APMG International APMG AI-Driven Project Manager (AIPM): AI governance, prompts, project controls, risks, and lifecycle decisions.

Exam Identity and Study Focus

Item	Reference
Vendor/provider	APMG International
Official exam title	APMG AI-Driven Project Manager (AIPM)
Official exam code	AIPM
Page purpose	Independent quick reference for candidates preparing for the real exam
Core mindset	Use AI to improve project outcomes while maintaining governance, accountability, ethics, assurance, and human judgment

The AIPM candidate should be ready to reason about how AI changes project management work, not merely define AI terms. Expect decision scenarios involving when to use AI, when not to use it, how to govern outputs, how to manage risk, and how to keep the project manager accountable.

High-Yield Exam Lens

If the question asks about…	Think first about…	Common trap
Using AI to make a project decision	Human accountability, evidence, context, stakeholder impact	Treating AI output as automatically correct
Automating project work	Suitability, risk, data quality, control points	Automating a poor process
AI-generated plans, estimates, or reports	Validation, assumptions, traceability	Presenting generated content without review
Sensitive project data	Privacy, confidentiality, access control, data minimization	Pasting restricted data into unmanaged tools
Bias or unfair outcomes	Dataset bias, model behavior, impacted stakeholders	Assuming “algorithmic” means objective
Predictive analytics	Historical data quality, uncertainty, confidence, explainability	Confusing prediction with certainty
AI governance	Policy, roles, approvals, auditability, escalation	Treating governance as an IT-only concern
Benefits realization	Measurable value, adoption, behavioral change	Measuring tool deployment instead of outcomes
Agile or hybrid delivery	Continuous feedback, experimentation, transparency	Using AI to hide uncertainty from stakeholders

AI-Driven Project Manager Core Responsibilities

Responsibility	What it means in AIPM-style scenarios
Select appropriate AI use cases	Choose AI where it improves decision quality, speed, insight, consistency, or productivity without unacceptable risk
Maintain human accountability	AI may assist, recommend, summarize, classify, or forecast; the accountable project role still owns decisions
Govern data and outputs	Protect sensitive data, manage access, validate outputs, keep records where needed
Challenge AI recommendations	Check assumptions, ask for sources or rationale, compare with project evidence, involve experts
Manage AI-related risk	Identify risks from data quality, bias, hallucination, security, overreliance, integration, compliance, and adoption
Adapt project processes	Embed AI into planning, monitoring, reporting, stakeholder engagement, lessons learned, and benefits tracking
Build AI literacy	Help the team understand proper use, limitations, escalation paths, and ethical expectations
Measure value	Track whether AI improves project outcomes, not just whether the tool is being used

AI Concepts for Project Managers

Term	Project-management meaning	Exam distinction
Artificial intelligence	Systems performing tasks that normally require human intelligence	Broad umbrella; not all AI is generative AI
Machine learning	Models that learn patterns from data	Output quality depends heavily on training and input data
Generative AI	AI that creates text, images, plans, summaries, code, or other content	Useful for drafting; requires validation
Large language model	Model trained to predict and generate language	Can be fluent and wrong
Prompt	Instruction or input given to an AI system	Prompt quality strongly affects output quality
Hallucination	Plausible but incorrect or unsupported AI output	Must be controlled by verification
Bias	Systematic unfairness or distortion in input data, design, or output	Can affect prioritization, decisions, and stakeholder treatment
Explainability	Ability to understand why a model produced an output	More important for high-impact decisions
Automation	System executes a task with limited human involvement	Needs controls, monitoring, and fallback
Augmentation	AI supports human work without replacing judgment	Often the safer default for project management
Predictive analytics	Uses data to forecast likely outcomes	Forecasts are probabilistic, not guarantees
Natural language processing	AI processing human language	Relevant to document analysis, meeting notes, sentiment, requirements
Computer vision	AI interpreting images or video	Relevant in construction, quality inspection, asset monitoring
Digital twin	Digital representation of a system or asset	Useful for simulation, scenario testing, and operational planning

AI Use Cases Across the Project Lifecycle

Lifecycle area	AI can help with	Project manager must ensure
Business case	Market scanning, option comparison, benefit hypothesis drafting	Assumptions, strategic alignment, value logic, sponsor validation
Initiation	Stakeholder mapping, charter drafting, lessons learned retrieval	Correct context, authority, objectives, constraints
Planning	Work breakdown suggestions, schedule drafts, risk identification, estimate ranges	Team review, dependency logic, realistic assumptions
Estimating	Analogous data analysis, effort forecasting, uncertainty ranges	Data relevance, expert challenge, contingency rationale
Risk management	Risk pattern detection, response suggestions, early-warning indicators	Ownership, probability/impact assessment, response feasibility
Stakeholder engagement	Sentiment analysis, communication tailoring, message drafting	Tone, accuracy, inclusion, confidentiality
Procurement	Supplier research, requirement drafting, bid comparison support	Fairness, transparency, conflict of interest controls
Delivery monitoring	Anomaly detection, progress forecasts, issue clustering	Current data, escalation thresholds, corrective action
Reporting	Draft status reports, dashboards, variance explanations	Accuracy, materiality, audience needs, no hidden uncertainty
Change control	Impact analysis, option modeling, documentation support	Governance route, decision authority, baseline control
Quality	Defect pattern analysis, test prioritization, inspection support	Acceptance criteria, sampling limits, human verification
Benefits realization	Adoption signals, benefit tracking, outcome analytics	Benefits owner accountability, measurement integrity
Closure	Lessons learned summarization, document indexing, handover checklists	Completeness, knowledge retention, final acceptance

Use-Case Selection Matrix

Use case type	Good AI candidate when…	Avoid or tightly control when…	Preferred control
Summarization	Source material is available and low sensitivity	Nuance, legal meaning, or commitments may be lost	Human review against source
Drafting	Output is a first draft for expert refinement	Stakeholders may treat draft as approved	Mark as draft; approval workflow
Classification	Categories are clear and repeatable	Misclassification creates high impact	Sampling, audit, exception review
Prediction	Historical data is relevant and sufficiently reliable	Novel project, sparse data, major context change	Confidence ranges and expert challenge
Recommendation	Decision criteria are known	Ethical, contractual, safety, or strategic consequences are high	Decision log with rationale
Automation	Task is repetitive, rules-based, low ambiguity	Exceptions are frequent or costly	Human-in-the-loop and rollback
Monitoring	Data streams are timely and meaningful	False positives or false negatives cause harm	Threshold tuning and escalation path
Stakeholder sentiment	Large volumes of text need pattern detection	Small sample, sensitive HR context, cultural nuance	Aggregate analysis; avoid individual profiling

“What Should the Project Manager Do Next?” Decision Table

Scenario	Best next action	Why
AI tool produces a schedule that looks optimistic	Validate assumptions with team and compare to historical data	AI output is an input, not an approved baseline
Sponsor asks to use public AI with confidential project documents	Check organizational policy and data classification before use	Protect confidentiality and comply with governance
AI identifies a high-risk supplier pattern	Investigate evidence and engage procurement/risk owners	Avoid acting on unverified AI conclusions
Team wants to automate status reporting	Define data sources, review process, exception handling, and accountability	Automation must preserve accuracy and ownership
AI-generated estimate conflicts with expert estimate	Examine assumptions, data relevance, and uncertainty; reconcile transparently	Conflicting evidence should improve estimate quality
Stakeholders are concerned AI will replace roles	Communicate purpose, controls, impacts, and involvement plan	Adoption depends on trust and transparency
AI model performance degrades during delivery	Pause or limit reliance, investigate data/model changes, escalate	Model drift can undermine decisions
Generated requirements contain ambiguity	Facilitate stakeholder clarification and acceptance criteria definition	AI can draft, but stakeholders define needs
AI recommends cancelling a workstream	Treat as decision support; assess business case, risks, dependencies, and governance	Major changes require authorized decision-making
AI output cannot explain its reasoning	Increase human review or use a more explainable method for high-impact use	Explainability matters when consequences are significant

Governance Reference

AI Governance Objectives

Objective	Practical meaning for projects
Accountability	Named people remain responsible for decisions and outcomes
Transparency	Stakeholders understand where AI is used and why
Fairness	Outputs are checked for bias or disproportionate impact
Privacy	Personal and sensitive data are protected and minimized
Security	Tools, integrations, and data flows are controlled
Quality	Outputs are validated before use
Traceability	Important AI-assisted decisions can be reconstructed
Compliance	Organizational policy, contracts, and applicable obligations are followed
Value	AI use is justified by measurable project or business benefit

Governance Controls by Risk Level

AI use risk	Examples	Suitable controls
Low	Drafting meeting agenda, summarizing non-sensitive notes	User review, prompt hygiene, version control
Medium	Risk identification, stakeholder communication drafts, schedule suggestions	Peer review, source validation, decision log
High	Supplier scoring, project funding recommendations, safety-related analysis	Formal approval, explainability, audit trail, expert review
Very high	Decisions affecting employment, legal rights, safety, regulated outcomes	Avoid unless explicitly authorized and strongly controlled

Human-in-the-Loop Patterns

Pattern	Description	Best used for
Human-in-the-loop	Human reviews before output is used	Reports, estimates, communications
Human-on-the-loop	AI operates but human monitors and can intervene	Dashboards, alerts, workflow routing
Human-in-command	Human sets objectives, constraints, approvals, and escalation rules	High-impact project decisions
Full automation	AI/system acts without routine human intervention	Low-risk, repeatable tasks with clear rules

For exam scenarios, prefer augmentation with accountable human oversight unless the task is low-risk, repeatable, and well controlled.

Data Management and AI Quality

Data issue	Effect on AI-enabled project work	Candidate response
Incomplete data	Missed risks, weak forecasts, false confidence	Identify gaps and qualify conclusions
Outdated data	Forecasts reflect past conditions, not current reality	Refresh sources and check assumptions
Biased data	Unfair or distorted recommendations	Test for bias; involve diverse review
Poorly labeled data	Weak classification or prediction accuracy	Improve data definitions and labels
Inconsistent definitions	Conflicting dashboards and reports	Establish common data dictionary
Sensitive data exposure	Privacy, confidentiality, contractual risk	Minimize, anonymize, or use approved tools
Lack of provenance	Cannot verify source or reliability	Require source traceability
Data drift	Model performance worsens over time	Monitor performance and recalibrate

Prompt Engineering for Project Managers

Practical Prompt Structure

Use prompts that define the role, objective, context, inputs, constraints, output format, and validation request.

Role: Act as a project controls analyst.
Objective: Identify schedule risks in the following status data.
Context: The project is in execution; baseline dates must not be changed without approval.
Inputs: [paste approved, non-sensitive data]
Constraints: Do not invent missing data. Flag assumptions separately.
Output: Provide a table with risk, evidence, likely impact, owner, and suggested next action.
Validation: List any data gaps or uncertainties.

Prompt Patterns

Pattern	Use when	Example instruction
Role framing	You need domain-specific structure	“Act as a project risk facilitator…”
Context grounding	Output must fit the project environment	“Use the approved scope and constraints below…”
Source-bound response	Accuracy matters	“Use only the provided text; do not infer missing facts.”
Assumption listing	Inputs are incomplete	“Separate facts, assumptions, and open questions.”
Comparative analysis	Options must be evaluated	“Compare options using cost, time, risk, and benefit.”
Critique mode	You need challenge, not agreement	“Identify weaknesses in this plan.”
Scenario testing	You need impact analysis	“Assess effects if supplier delivery slips by four weeks.”
Output formatting	You need usable artifacts	“Return a risk register table with owner and response.”

Prompt Quality Checklist

Check	Question
Purpose	What decision or artifact will this support?
Data	Is the input approved, current, and safe to use?
Context	Have constraints, lifecycle, stakeholders, and assumptions been stated?
Boundaries	Have you told the AI what not to do?
Output	Is the format actionable for the project process?
Validation	Have you asked for gaps, uncertainty, and assumptions?
Review	Who will check the output before use?

AI Output Validation

Validation method	Use for	What to check
Source comparison	Summaries, requirements, decisions	Does output match the source?
Expert review	Estimates, risks, solution options	Is it realistic and context-aware?
Data reconciliation	Dashboards, forecasts, reports	Does it match approved systems of record?
Sensitivity analysis	Forecasts and scenarios	How do results change when assumptions change?
Bias review	Stakeholder, supplier, or people-related analysis	Are outcomes unfairly skewed?
Red-team challenge	Important plans or recommendations	What could be wrong, missing, or manipulated?
Pilot testing	New AI workflow	Does it work safely before scaling?
Audit trail	Material AI-assisted decisions	Can the reasoning and inputs be reconstructed?

Risk Management for AI-Driven Projects

AI-Specific Risk Register Examples

Risk	Cause	Possible impact	Response options
Hallucinated project information	Generative AI invents facts	Wrong reports, decisions, commitments	Source-bound prompts, review, citations
Data leakage	Sensitive data entered into unapproved tool	Confidentiality breach	Approved tools, data classification, training
Biased recommendations	Skewed historical data or model design	Unfair supplier or stakeholder treatment	Bias testing, human review, diverse input
Overreliance on AI	Team accepts outputs without challenge	Poor decisions, loss of expertise	Accountability rules, review checklists
Model drift	Conditions change after model design	Forecasts become unreliable	Monitor accuracy, recalibrate, fallback
Lack of explainability	Black-box recommendation	Weak trust, poor governance	Require rationale, use explainable tools
Integration failure	AI tool not aligned with project systems	Duplicate data, errors, rework	Architecture review, controlled rollout
Cybersecurity exposure	New APIs, plugins, or data flows	Unauthorized access or manipulation	Security assessment, access control
Poor adoption	Users distrust or misunderstand AI	Benefits not realized	Training, communication, involvement
Automation of flawed process	Inefficient process is accelerated	Faster errors and waste	Improve process before automation

Risk Response Selection

Response	When appropriate	AI-related example
Avoid	Risk is unacceptable and value is low	Do not use public AI for restricted data
Reduce/mitigate	Risk can be lowered with controls	Add human review before AI-generated reports
Transfer/share	Another party is better placed to manage part of the risk	Contractual support from approved AI vendor
Accept	Risk is tolerable and monitored	Use AI drafting for low-impact internal notes
Escalate	Risk exceeds project manager authority	AI use may affect legal, regulatory, or enterprise policy issues

Ethics and Responsible AI

Principle	Project-management application
Human agency	People remain able to challenge, override, and decide
Transparency	Disclose material AI use where relevant
Fairness	Check for discriminatory or exclusionary outcomes
Privacy	Use the minimum necessary personal data
Security	Protect prompts, outputs, integrations, and stored data
Reliability	Validate before acting; monitor performance
Accountability	Assign owners for AI use, review, and outcomes
Proportionality	Match governance effort to risk and impact

Ethical Decision Traps

Trap	Better exam answer
“The AI recommended it, so we should proceed.”	Treat recommendation as evidence; validate and decide through governance
“We can use any data because it improves accuracy.”	Use only appropriate, authorized, minimized data
“Bias is only a technical problem.”	Bias is also governance, stakeholder, and decision-quality risk
“Transparency means exposing all technical details.”	Provide understandable explanation appropriate to the audience
“Human review always solves the issue.”	Review must be competent, independent where needed, and evidence-based

Project Controls and AI

Earned Value and Forecasting Essentials

AI may support variance explanation and forecasting, but the project manager must understand the control logic.

\[ \text{Cost Variance (CV)} = \text{Earned Value (EV)} - \text{Actual Cost (AC)} \]\[ \text{Schedule Variance (SV)} = \text{Earned Value (EV)} - \text{Planned Value (PV)} \]\[ \text{Cost Performance Index (CPI)} = \frac{\text{EV}}{\text{AC}} \]\[ \text{Schedule Performance Index (SPI)} = \frac{\text{EV}}{\text{PV}} \]

Indicator	Plain meaning	Typical interpretation
CV	Value earned minus cost spent	Negative means over budget
SV	Value earned minus value planned	Negative means behind planned progress
CPI	Cost efficiency	Less than 1.0 means poor cost efficiency
SPI	Schedule efficiency	Less than 1.0 means poor schedule efficiency

AI Support for Controls

Control activity	AI contribution	PM caution
Variance analysis	Identify patterns and likely drivers	Confirm with actual project evidence
Forecasting	Predict completion trends	Use ranges; do not hide uncertainty
Dashboarding	Generate summaries and alerts	Validate data sources and thresholds
Corrective actions	Suggest options	Assess feasibility, authority, and side effects
Lessons learned	Cluster recurring issues	Avoid losing context or dissenting views

Estimation and Prioritization

Technique	AI can support by…	Watch for…
Analogous estimating	Searching comparable past projects	False similarity
Parametric estimating	Applying historical relationships	Poor calibration
Three-point estimating	Structuring optimistic, most likely, pessimistic values	Unrealistic ranges
Monte Carlo-style simulation	Exploring probability distributions	Invalid assumptions and weak input data
MoSCoW prioritization	Grouping requirements	Stakeholder authority and value logic
Weighted scoring	Comparing options against criteria	Hidden bias in weights
Cost-benefit analysis	Drafting benefit and cost categories	Unverified benefit claims

Three-point expected value is commonly expressed as:

\[ \text{Expected Value} = \frac{O + 4M + P}{6} \]

Where \(O\) is optimistic, \(M\) is most likely, and \(P\) is pessimistic.

Change Control in AI-Enabled Projects

Change situation	AI may help with	Governance action
Scope change request	Impact analysis, dependency discovery, document drafting	Submit through agreed change control
AI tool change	Benefit/risk comparison, implementation checklist	Assess security, data, process, training impacts
Model or configuration update	Release notes, test scenario generation	Retest outputs and update controls
Baseline impact	Forecast schedule/cost effects	Obtain authorized approval before rebaselining
Stakeholder impact	Communication drafts and sentiment analysis	Confirm messages and engagement plan

Change Decision Logic

    flowchart TD
	    A[Proposed AI or project change] --> B{Affects scope, cost, schedule, risk, quality, benefits, or governance?}
	    B -- No --> C[Handle within team authority and record if useful]
	    B -- Yes --> D{Within project manager tolerance?}
	    D -- Yes --> E[Assess impact, consult owners, approve per delegated authority]
	    D -- No --> F[Escalate to appropriate governance body]
	    E --> G[Update plans, logs, controls, and communications]
	    F --> G

Stakeholder Engagement with AI

Engagement task	AI use	Required human judgment
Stakeholder identification	Suggest stakeholder groups from documents	Confirm influence, interest, legitimacy
Sentiment analysis	Detect patterns in feedback	Interpret culture, context, and sample limits
Communication planning	Tailor messages by audience	Ensure accuracy, empathy, and transparency
Meeting support	Summaries, actions, decisions	Validate commitments and owners
Conflict analysis	Identify themes and concerns	Facilitate resolution directly
Training and adoption	Generate learning materials	Address role impact and resistance

Communication Rules for AI Use

Rule	Application
Be transparent where material	Tell stakeholders when AI materially shapes outputs or processes
Avoid false precision	Present forecasts as ranges or scenarios when uncertain
Do not outsource accountability	Project manager or owner signs off
Protect sensitive content	Use approved channels and data handling
Keep messages human	AI can draft; people manage trust

Agile, Predictive, and Hybrid Considerations

Delivery approach	AI fits well for	Watch for
Predictive	Planning, baseline analysis, documentation, forecasting	Overconfidence in early AI-generated plans
Agile	Backlog refinement, user story drafting, test generation, feedback analysis	AI-generated stories without user validation
Hybrid	Roadmap planning, dependency analysis, reporting across teams	Conflicting governance cadences
Product-focused	Usage analytics, experiment analysis, feature prioritization	Measuring output rather than outcome
Programme/portfolio context	Scenario modeling, dependency mapping, investment options	Weak comparability across data sets

Agile AI Traps

Trap	Better response
AI writes user stories without product owner input	Use AI drafts; product owner validates value and acceptance criteria
AI prioritizes backlog solely by volume of requests	Combine analytics with strategy, value, risk, and stakeholder judgment
AI-generated velocity forecasts are treated as commitments	Use forecasts to support planning, not to pressure teams
Retrospectives are summarized without psychological safety	Validate themes and preserve trust

Procurement and Supplier Management

Area	AIPM-relevant focus
Supplier AI claims	Validate capability, evidence, assumptions, and limitations
Data ownership	Clarify who can access, store, train on, or reuse project data
Security	Assess integration, access control, audit, and incident handling
Service levels	Define performance, availability, support, and escalation expectations
Exit strategy	Avoid lock-in; plan data export and continuity
Evaluation fairness	Use consistent criteria; avoid opaque or biased scoring
Contract change	Treat AI tool changes as potential changes to risk, cost, process, and obligations

Benefits and Value Realization

Benefit type	Possible AI-enabled measure	Caution
Productivity	Time saved drafting, summarizing, searching	Avoid counting gross time saved without quality check
Decision quality	Fewer late surprises, improved forecast accuracy	Need baseline and comparable measures
Risk reduction	Earlier detection of issues	Track response effectiveness, not just alerts
Stakeholder experience	Faster, clearer communication	Monitor trust and satisfaction
Quality	Reduced defects or rework	Ensure AI is not masking root causes
Knowledge management	Faster retrieval of lessons and documents	Keep sources current and governed

Benefits Logic

Level	Example
Input	AI tool, data, training
Activity	Automated summary and risk pattern detection
Output	Faster reports and earlier alerts
Outcome	Better-informed decisions and fewer unmanaged risks
Benefit	Reduced delay, cost avoidance, improved delivery confidence

Assurance, Auditability, and Documentation

Artifact	What to capture
AI use register	Where AI is used, purpose, owner, risk level, tool, controls
Data classification record	What data is used and whether it is approved for the tool
Prompt/output record	For material decisions, keep key prompts, inputs, outputs, and versions
Decision log	Human decision, rationale, evidence, and AI contribution
Risk register	AI-specific risks and controls
Validation checklist	Review method and reviewer
Change record	Approved changes to AI tools, workflows, models, or controls
Lessons learned	What worked, what failed, and reusable guidance

Common Scenario Distinctions

Distinction	Choose this when…	Not this when…
AI as advisor	Decision is complex and needs human accountability	You need a simple, approved rules-based workflow
AI as automation	Task is repeatable, low-risk, and measurable	Task needs judgment, empathy, or escalation
Public AI tool	Data is non-sensitive and policy allows	Data is confidential, personal, contractual, or restricted
Approved enterprise AI	Governance, security, and data controls are required	Tool has not been assessed or authorized
Explainable model	Decisions are high-impact or need stakeholder trust	Output is low-impact drafting
Generative AI	Need drafts, summaries, scenarios, or language support	Need guaranteed factual correctness without validation
Predictive model	Historical data is meaningful	Project is novel or data is weak
Dashboard alert	Need early warning	Alert thresholds are untested or noisy
Full automation	Low-risk process with clear rules	Exceptions require judgment

Quick Checklists

Before Using AI on a Project Task

Is the task suitable for AI support?
Is the tool approved for this data and purpose?
Is the data accurate, current, and appropriately classified?
Have assumptions and constraints been stated?
Is there a human owner for review and decision?
Is the output risk level understood?
Is a record needed for audit or decision traceability?
Are stakeholders affected or required to be informed?

Before Trusting an AI Output

Does it match source evidence?
Are assumptions clearly separated from facts?
Are gaps and uncertainties identified?
Does it align with project objectives, constraints, and governance?
Has an appropriate expert or owner reviewed it?
Could bias, missing context, or outdated data affect it?
Is the recommendation proportionate to the evidence?
Is escalation needed before action?

Before Scaling an AI Practice

Pilot first with defined success criteria.
Measure quality, time, risk, and adoption.
Document controls and ownership.
Train users on appropriate use and limitations.
Monitor unintended consequences.
Update governance, workflows, and lessons learned.

Fast Exam Traps to Avoid

Trap	Better AIPM exam response
Replacing project governance with AI recommendation	Use AI within governance, not instead of it
Assuming AI removes the need for stakeholder engagement	AI may support engagement; it cannot replace trust-building
Using more data without considering permission or relevance	Use appropriate, authorized, quality data
Treating AI predictions as commitments	Present uncertainty and validate with experts
Ignoring organizational policy	Check approved tools, data rules, and escalation paths
Automating a weak process	Improve or clarify the process first
Over-focusing on tool features	Focus on outcomes, risk, value, and control
Hiding AI use from affected stakeholders	Be transparent where use is material
Trusting fluent language	Verify facts, assumptions, and sources
Making AI an IT-only issue	Treat AI as a project, governance, people, and value issue

Final Review Map

Topic	Candidate should be able to answer
AI fundamentals	What type of AI is being used and what are its limitations?
Project lifecycle	Where can AI improve planning, delivery, monitoring, and closure?
Governance	Who is accountable and what controls are required?
Data	Is the data safe, relevant, accurate, and authorized?
Risk	What new risks does AI introduce and how are they managed?
Ethics	Are fairness, transparency, privacy, and human agency protected?
Change	Does AI use require impact assessment or formal approval?
Stakeholders	How will AI affect trust, communication, adoption, and roles?
Benefits	What measurable value is expected and how will it be proven?
Assurance	Can important AI-assisted decisions be reviewed later?

Practical Next Step

Use this Quick Reference to build short scenario drills: for each AI use case, decide whether to use AI, what controls are needed, who remains accountable, what risks arise, and what the project manager should do next. Then move into timed AIPM-style practice questions to test decision-making under exam conditions.

Scenario Guide

Embracing AI in Project Management and Basic Concepts