PMI-CPMAI
Free PMI Certified Professional in Managing AI Practice Test
12 exam-style questions with full explanations — no sign-up. Score yourself, then close your gaps with the full course.
Exam fee ~$6995 exam domainsLevel Intermediate2 timed practice tests in the course
✅ Free practice — no sign-up📝 Real exam-style questions💡 Detailed explanations💸 30-day money-back via Udemy
A healthcare AI vendor is building a model to predict patient readmission risk using historical EHR data that includes protected health information. Before data collection begins, the project team wants to identify and mitigate privacy risks to individuals whose data will be used, and document the legal basis for processing. Which activity BEST accomplishes this within the CPMAI Responsible AI practices?
A Privacy Impact Assessment (PIA) is the formal, documented process for identifying privacy risks to individuals, evaluating necessity/proportionality of processing, and defining mitigations before data use begins — a core CPMAI responsible AI practice tied to GDPR/CCPA compliance.
A retail company wants an AI system that automatically routes incoming customer emails into categories like 'billing', 'shipping', and 'returns' based on the text content of each message. Which of the Seven Patterns of AI best matches this business need?
Classifying unstructured text into predefined categories based on recognizing patterns in the content (words, phrases, structure) is the core definition of the Recognition Pattern, which covers recognizing patterns in images, sound, and text — including text classification tasks like this one.
A company is building a predictive maintenance AI model. During the Identify Data Needs phase, the team discovers sensor data covers only 60% of required equipment types, spans 3 years of history, but has 40% missing values in critical readings. Per the CPMAI methodology, what is the BEST next step?
CPMAI's go/no-go gate discipline requires that when data does not meet the volume, coverage, or quality needed for the solution, the team issue a no-go decision and cycle back to reassess business objectives or find additional/alternate data sources rather than pushing forward on unfit data.
A healthcare AI project team has completed data preparation and wants to proceed into model development. Which of the following BEST demonstrates that the data quality go/no-go gate should be passed?
The CPMAI data preparation go/no-go gate requires documented evidence that data quality dimensions meet thresholds appropriate to the use case before proceeding to modeling.
An AI model for credit risk scoring has been in production for eight months. The monitoring dashboard shows the model's prediction accuracy has dropped from 92% to 81% over the last six weeks, while the underlying data pipeline shows no errors. What should the operations team do FIRST?
A gradual accuracy decline with no pipeline errors is a classic signature of drift; the correct first step in production governance is to diagnose the type and cause of drift (data drift vs. concept drift) before deciding on remediation such as retraining or rollback.
During model evaluation, a hiring-recommendation model shows a 78% approval rate for one demographic group compared to 52% for another with similar qualifications, despite race and gender being excluded as direct model features. What is the MOST likely explanation the team should investigate first?
Bias checks must examine not just direct protected attributes but proxy variables (e.g., zip code, school name) that correlate with them; this is a well-documented cause of disparate outcomes even when protected fields are removed, and CPMAI requires bias analysis across data, model, and algorithm.
During the Identify phase, a project team wants to build a fraud-detection model but discovers only 500 labeled fraud examples exist across three years of transaction history. According to CPMAI feasibility evaluation practices, what should the team recommend?
CPMAI feasibility evaluation requires honestly comparing AI against traditional alternatives; when data volume and quality are insufficient for reliable model training, a traditional rule-based approach paired with a plan to build the labeled dataset over time is the responsible recommendation rather than forcing an AI solution prematurely.
An AI team needs authoritative definitions of the 'customer churn' fields stored in the CRM system, along with knowledge of historical schema changes affecting those fields. Which role should the team engage during the Identify Data Needs phase?
CPMAI explicitly calls for identifying data SMEs and data stewards who hold authoritative knowledge of field definitions, schema history, and data lineage within source systems like a CRM — critical input for scoping data requirements accurately.
A fraud detection model achieves 96% accuracy on training data but only 71% on the holdout test set, with performance dropping further on a separate out-of-time validation set. What should the team recommend at the operationalization go/no-go gate?
The operationalization gate requires evidence of robustness and generalization, not just training performance; the large train/test/out-of-time gap indicates overfitting that must be addressed before deployment.
During go-live of a new fraud detection model, the deployment team discovers that the production scoring API is returning malformed responses for 15% of transactions, causing downstream order processing failures. According to CPMAI deployment planning best practices, what should have been prepared in advance to handle this exact situation?
Deployment planning under CPMAI explicitly requires a rollback and contingency procedure with clear triggers (e.g., error rate thresholds) and a validated fallback so the team can quickly revert or mitigate when production issues occur, minimizing business disruption.
A regulator asks an AI project team to explain why a loan-denial decision was made by their machine learning model for a specific applicant. Which capability should the team have built into the AI system from the outset to satisfy this request?
Per-decision explainability tools (e.g., SHAP/LIME-style feature attribution) directly answer 'why did the model decide this for this individual,' which is the AI/ML transparency requirement CPMAI emphasizes for auditable algorithmic decisions.
An HR department wants to deploy an AI tool that screens job applicant resumes and ranks candidates before human review. As part of the Identify phase risk assessment, what should the project team prioritize before approving this project to move forward?
CPMAI's Identify phase requires risk assessment across security, safety, AND ethics before proceeding. For an AI system influencing hiring decisions, evaluating bias and disparate impact potential — and documenting mitigations — is essential given the high-stakes, legally sensitive nature of employment decisions.
Ready for the real thing?
The full course: two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed explanations.
$109.99 $34.99 with code FREETEST33 — valid through September 14.
Get my $34.99 deal →PMI-CPMAI exam — quick answers
How much does the PMI-CPMAI exam cost?
The exam fee is approximately $699 and varies by region — confirm current pricing with the certification vendor before you book.
What topics are on the exam?
It covers 5 domains: Support Responsible and Trustworthy AI Efforts (15%), Identify Business Needs and Solutions (26%), Identify Data Needs (26%), Manage AI Model Development and Evaluation (16%), Operationalize AI Solution (17%). The full course has a dedicated chapter, lab and practice-test coverage for each.
Is this practice test really free?
Yes — all questions on this page are free with explanations and no sign-up. The paid Udemy course adds two full-length timed exams, video lessons and hands-on labs.
How do I get the discount?
Use code FREETEST33 at checkout for $34.99 (list $109.99) through September 14 — the enroll button applies it automatically.
Will this prepare me for the real exam?
The questions mirror the real exam's style and are mapped to the official domains. This is exam-focused preparation — combine the free test with the full course's timed simulations to gauge your readiness.