During a process redesign initiative at a large distribution operation, a finance workflow is evaluated for possible automation. The activity supports a very high transaction volume each month and follows standardized validation steps tied to upstream procurement records. While the process operates within clearly defined rules, it also includes escalation thresholds for mismatches and periodic audit sampling to ensure compliance with internal controls. Using the Task Allocation Matrix, how should the automation potential of this task be categorized?

• A. Human-led Strategy
• B. Collaborative Interpretation
• C. Human Negotiation
• D. Full automation potential

As the AI Program Manager, you have completed the initial data collection for an enterprise AI readiness assessment. During the assessment review, you notice that the IT and Operations departments hold conflicting views regarding who should own data governance, leading to a stalemate. You need to move beyond individual data collection and bring these cross-functional teams together in a shared setting to openly discuss the findings, surface differing perspectives, and collectively agree on the priority issues. Which specific assessment technique is defined by its ability to build consensus and create shared ownership of next steps?

• A. Gap Analysis
• B. Surveys
• C. Workshops
• D. Heat Maps

A new predictive maintenance system was deployed on the factory floor three months ago. Despite technical validation confirming the model's accuracy, utilization reports show zero engagement. Shift supervisors report that their teams are reverting to legacy manual checklists because they cannot bridge the gap between the system's probabilistic dashboards and their standard operating procedures. Which specific adoption challenge is the primary cause of this project's stagnation?

• A. Skill Gap and Workforce Adaptation
• B. Ethical and Societal Risks
• C. Regulatory Compliance and Governance
• D. Human-AI Collaboration

During an AI initiative review, a delivery team reports that a predictive model is underperforming despite using datasets that already meet established quality, completeness, and consistency standards. The data has been sourced and validated, and no changes to model design or additional data acquisition are planned at this stage. Analysis indicates that existing data fields do not sufficiently reflect higher-level business behavior needed for learning. As part of AI operations oversight, you are asked to identify which data preparation activity should be applied next to address this issue. Which activity within the Data Collection and Preparation phase directly supports improving how existing data is represented for model learning?

• A. Creating meaningful variables from existing data
• B. Extracting raw data from source systems
• C. Applying ground truth labels to records
• D. Dividing data into training, validation, and test sets