An online banking platform uses an AI virtual assistant as its primary customer support channel. The project manager must prepare a business continuity plan for when the AI system goes offline for maintenance.Which strategy best ensures service continuity during downtime?
Answer(s): C
PMI-CPMAI risk and resilience practices emphasize graceful degradation: when the primary AI system is unavailable, customers must still receive some level of service. An automated fallback chatbot provides a simplified but always-available channel, maintaining a consistent experience and avoiding a complete service outage. This reflects the 'fail-soft' strategy in AI operations and MLOps: the system automatically routes to a simpler, more reliable backup. Logging helps with diagnosis but does not serve customers. Manual escalation and staff training are human-dependent and slower to activate than an automated fallback.
A fraud detection model that previously achieved strong accuracy has recently begun producing a rising number of incorrect predictions. No changes have been made to the model's architecture or training pipeline.What should the team investigate first to identify the root cause?
Answer(s): D
PMI-CPMAI monitoring guidance establishes that when a previously high-performing model shows unexpected error increases, the first diagnostic step is to examine production data distributions for drift. Changes in user behavior, population characteristics, upstream system outputs, or environmental conditions often cause models to degrade without any code change. Teams should compare current production feature distributions against training baselines before considering architectural fixes, preprocessing corrections, or regularization changes. Data drift analysis is the primary root-cause investigation tool in PMI's AI lifecycle framework.
A management consulting firm is advising a manufacturing client on whether to invest in AI for demand forecasting. The project manager needs to build a clear justification for choosing AI over conventional statistical methods.What is the most effective way to support the AI recommendation?
Answer(s): B
PMI-CPMAI frames the AI vs. non-AI decision as a business case question, not a technology preference. The framework requires project leaders to compare AI-based and non-AI alternatives using structured cost-benefit analysis that quantifies implementation costs, operational costs, expected value, and non-financial impacts such as risk and compliance. AI should only be selected when it provides clear incremental value over simpler options. Emphasizing novelty, trends, or complexity without quantified comparison does not satisfy the governance standard.
A national health insurer is deploying an AI system to coordinate care across 40 hospital partners in 12 countries. The training dataset must be representative of all patient populations served.Which step is most critical to optimize the training data?
PMI-CPMAI treats data understanding and preparation as the foundational step before any bias detection or KPI refinement. The guidance calls for profiling data sources to assess completeness, consistency, and coverage across relevant populations, then cleaning, integrating, and transforming data into a representative, analysis-ready dataset. In a multi-country healthcare scenario this includes schema harmonization, missing-value handling, and ensuring all relevant demographics and care pathways are adequately represented. Simply increasing dataset size without ensuring diversity can reinforce existing biases.
A telecom company is preparing training data for a network fault prediction model. The dataset combines structured telemetry logs and unstructured technician notes. The team needs everything in a consistent, model-ready format.What method best addresses this requirement?
PMI-CPMAI's data preparation guidance emphasizes that when projects work with both structured and unstructured data, standardization and transformation pipelines are required so downstream models receive well-defined schemas, consistent encodings, and unified representations. A data transformation tool fulfills this role: it normalizes formats, unifies categorical encodings, aligns date and unit standards, and converts unstructured content into structured features or embeddings, all within repeatable, auditable pipelines. Converting unstructured data alone is only one part of the broader standardization need.
A consumer goods manufacturer is operationalizing an AI quality control system on its assembly line. The project manager is asked to establish the initial data privacy and compliance framework.What is the most effective initial approach?
PMI's AI governance perspective treats data privacy and compliance as part of a broader governance capability rather than isolated technical controls. A comprehensive data governance plan defines roles, responsibilities, policies, standards, and processes that ensure regulatory, contractual, and ethical obligations are met across the full lifecycle. It provides the basis for later specific mechanisms such as audits, encryption standards, and incident response. Implementing only one control (audits, encryption, or a task force) without an overarching governance structure leaves critical gaps.
An IT services provider is about to start preparing data for a predictive server-downtime model. The project manager must decide whether the available telemetry data is suitable before investing in data engineering work.Which technique is most appropriate for this decision?
PMI-CPMAI establishes that data quality assessment must precede data preparation. EDA is the recommended technique at this stage: it allows the team to examine distributions, detect missing values, outliers, noise, inconsistencies, and potential bias, and to confirm whether the available logs contain sufficient signal and consistent labeling to support reliable modeling. Augmentation and labeling are downstream enhancement techniques; cost-benefit analysis is a management tool, not a data quality method.
A cross-functional AI project team is preparing to move from data preparation into model development. The project manager wants to ensure all transparency and explainability obligations are satisfied before proceeding.Which activity directly fulfills this requirement?
PMI-CPMAI identifies transparency and explainability as core responsible AI practices. Transparency requires that stakeholders can understand how and why the AI system produces its outputs. The most direct mechanism is thorough documentation of the model's decision-making process, including model architecture, training data characteristics, feature importance, decision thresholds, known failure modes, and interpretability artifacts such as model cards. This documentation serves audits, risk reviews, and stakeholder communication. Data quality, ethics guidelines, and feedback mechanisms address other important aspects but do not directly establish explainability.
Post your Comments and Discuss PMI PMI-CPMAI exam prep with other Community members:
Our website is free, but we have to fight against AI bots and content theft. We're sorry for the inconvenience caused by these security measures. You can access the rest of the PMI-CPMAI content, but please register or login to continue.