PMI-CPMAI highlights data pipelines and preprocessing as critical components of AI/ML configuration management. A core principle is that all evaluation datasets must be processed through consistent, validated preprocessing steps (cleaning, normalization, feature engineering, encoding, etc.). If different test datasets experience different preprocessing logic, parameter settings, or transformations, performance metrics will naturally appear inconsistent, not because of the model itself but because the inputs are not comparable.
The guidance notes that configuration management for AI must track not only model versions but also data transformations, feature pipelines, and parameter settings. Inconsistent metrics across test datasets are a classic symptom of mismatched preprocessing, such as applying different scaling, missing-value handling, text tokenization, or feature selection strategies across datasets. Overfitting and model complexity affect generalization, but typically manifest as consistently poor performance on out-of-sample data, rather than erratic metrics between test sets prepared correctly.
Therefore, when a team observes inconsistent performance metrics across different test datasets, PMI-CPMAI would direct them to first check whether the data preprocessing steps are implemented correctly and consistently across those datasets. The likely cause of the inconsistency issue is incorrect (or inconsistent) data preprocessing steps.

In PMI-CPMAI-aligned practice, a go/no-go assessment is a formal checkpoint where technology, data, governance, risk, and stakeholder factors are evaluated against predefined criteria. If this assessment uncovers that multiple technology and data factors are insufficient, the appropriate response is not to proceed, but to pause and address those deficiencies. The project manager's role is to coordinate further analysis of data readiness (availability, quality, completeness, relevance) and verify that stakeholder expectations and commitments are still aligned with the AI initiative's constraints and risks.
Option A-verify data quality and stakeholder alignment-captures this corrective step. It reflects the PMI principle that AI projects must be based on trustworthy data and shared understanding; otherwise, model outcomes may be unreliable, non-compliant, or misaligned with business value. Options B, C, and D effectively ignore or downplay the red flags discovered in the assessment, which violates disciplined, risk-aware AI governance. Proceeding despite known gaps, focusing only on technology while neglecting data, or launching without further assessment directly contradicts structured go/no-go decision logic and could expose the organization to operational, ethical, or regulatory failure.
Therefore, the appropriate action after an unfavorable go/no-go outcome is to re-verify and remediate data quality issues and ensure stakeholder alignment (option A).

When an AI solution is described as "not scalable due to high maintenance requirements," PMI-style AI governance and lifecycle guidance points toward architectural refactoring rather than simply changing technologies or deployment environments. High maintenance often stems from tight coupling, monolithic design, and lack of clear separation between data, model, business logic, and interface layers.
Adopting a modular architecture to isolate different system components (option C) directly addresses this problem. In a modular or microservice-oriented design, each component-data ingestion, feature engineering, model training, model serving, monitoring, etc.-is separated behind clear interfaces. This makes it much easier to update or replace one part of the system without impacting the whole, which reduces maintenance overhead and improves scalability over time. It also supports independent deployment, targeted testing, and selective scaling of the components that receive the heaviest load.
Switching to a rule-based system (option A) typically increases maintenance complexity in dynamic environments. Incorporating generative AI (option B) may change the modeling approach but does not inherently solve structural maintenance issues. Utilizing cloud-based solutions (option D) helps with infrastructure scalability but does not fix architectural coupling. Therefore, the most effective way to address non-scalability caused by high maintenance requirements is to adopt a modular architecture.

PMI's Seven Patterns of AI describes Predictive analytics & decision support as using data-driven learning to anticipate outcomes and support decisions under uncertainty. Fraud detection is a classic predictive use case:
the system analyzes historical and current transaction behaviors to estimate the probability of fraud and recommend actions (approve, decline, escalate). In CPMAI-aligned delivery, the project manager ensures the AI capability matches the business objective and defines measurable performance metrics and thresholds (e.
g., false positives, fraud loss reduction, detection latency). PMI-CPMAI also emphasizes responsible and trustworthy AI practices-particularly around privacy, governance, and monitoring-because fraud models can affect customers' access to funds and may introduce bias if training data is skewed. Predictive analytics best fits because it supports classification/risk scoring decisions; the other options focus on interaction (conversational), tailored experiences (hyperpersonalization), or self-directed control (autonomous systems).

PMI guidance on prompts and prompt engineering states that prompts "supply the system with context and guidance as well as constraints," and that the value of a GenAI system "can only be realized through the instructions provided to it." PMI further explains that while an AI system can respond to very short inputs,
"the less specific a prompt, the more likely the results will be vague or unhelpful," explicitly linking insufficient specificity/context to degraded usefulness. In PMI's recommended "diverge and converge" prompt approach, teams begin broad, then progressively refine, adding details such as industry, region, project type, intended use, and examples-because "the granularity of the input will be directly proportional to the utility of the output received." Therefore, if the team provides insufficient context, the model must "guess" what is intended, which most directly manifests as answers that are not aligned to the actual task needs (i.e., lacking relevance), rather than being more accurate or more efficient.