PMI Certified Professional in Managing AI (PMI-CPMAI)

Learning plan

About 35 hours of focused study, split into 5 phases. Pick your weekly pace and the timeline adjusts.

Estimated finish: 4 weeks

Before you start

• No experience is required to sit the exam, and the fee includes the 21-hour CPMAI prep course
• Basic project-management vocabulary (scope, stakeholders, iterations). Agile experience helps because CPMAI is iterative

1. Week 1 · 6 h

   CPMAI methodology and responsible AI

   • Name the six CPMAI phases in order (Business Understanding, Data Understanding, Data Preparation, Model Development, Model Evaluation, Model Operationalization) and explain why it extends CRISP-DM
   • Explain why AI projects fail differently from software projects (data dependency, probabilistic outputs, model drift)
   • Name the seven patterns of AI (recognition, conversational, predictive analytics, hyperpersonalisation, autonomous systems, patterns and anomalies, goal-driven) and match them to use cases
   • Build trustworthy-AI checks (fairness, transparency, privacy, accountability) into every phase rather than leaving them to the end

   Use: PMI-CPMAI certification page · NIST AI Risk Management Framework

2. Weeks 1–2 · 8 h

   Business understanding and AI go/no-go

   • Turn a business problem into an AI problem statement with a measurable success metric and ROI hypothesis
   • Run a go/no-go check (business feasibility, data feasibility, execution feasibility) and decide when AI isn't the right tool
   • Scope a minimum viable AI deliverable and plan short iterations

   Hands-on: Pick a real process at work (for example invoice routing or churn prediction) and write a one-page CPMAI Phase I brief. Include the AI pattern, success metric, go/no-go answers and the smallest first iteration.

   Use: PMI-CPMAI certification page

3. Weeks 2–3 · 9 h

   Data understanding and data preparation

   • Inventory the data sources a model needs and assess quantity, quality, labelling, access rights and privacy constraints
   • Plan data preparation work (cleaning, labelling, augmentation, feature engineering) and estimate its share of project effort
   • Identify data-governance and bias risks before training starts

   Hands-on: Take a public dataset from Kaggle and write a data-understanding report covering record counts, missing values, label balance, sensitive attributes, licence and a list of the preparation steps needed. Try to finish in under two hours so you get a feel for how much effort data work takes.

   Use: Kaggle Learn – Data Cleaning

4. Week 3 · 7 h

   Model development, evaluation and operationalisation

   • Explain at a manager's level the trade-offs between building, fine-tuning and buying a model, and the main model families
   • Set evaluation criteria tied to the business metric, not just accuracy, and decide what happens if the model misses the threshold
   • Plan operationalisation, including deployment, monitoring for drift, retraining cadence, ownership and when to go back to an earlier phase

   Use: NIST AI RMF Playbook

5. Week 4 · 5 h

   Practice exams and review

   • Finish all module quizzes in the official prep course
   • Take at least one 120-question timed practice exam in 160 minutes and review every miss by CPMAI phase
   • Re-study the data phases, since data work is the core of the methodology

   Use: PMI-CPMAI certification page

You're ready to book when…

• You can list the six CPMAI phases and what is produced at the end of each, without notes
• You can run a go/no-go assessment on a new scenario in under five minutes
• You score 75% or higher on a full-length timed practice exam

Exam tips

• Scenario questions usually reward going back to an earlier phase (for example, back to Data Understanding when quality is poor) rather than pushing on.