The N.C. Department of Information Technology uses a set of evaluation criteria to determine if an AI use case aligns with the AI Accelerator’s goals, demonstrates potential for innovation and meets feasibility and ethical standards.

These criteria also help prioritize use cases that are not only feasible but also use cases that have a significant impact on the state.

Feasibility

Data Availability

• Is the data available and accessible?
• What is the data classification? (All data in the initial phase of the AI Accelerator must be “low risk,” publicly available data.)
• Are there data privacy or security concerns with the data?
• Is the data centralized or distributed across divisions?
• Are there known gaps in the data?
• Is this an ethical use of the data?

Resources

• Does the agency that owns the data have the technical and business expertise to implement and maintain the AI tool?
• Are there existing APIs or data pipelines?
• Can the agency dedicate personnel to support the AI Accelerator?
• Can the agency commit to weekly status meetings throughout the engagement?

Timelines

• Can the pilot be completed within 60 days?

Impact

Magnitude

• What is the potential scope?
  • Individual/Team
  • Division
  • Agency
  • Inter-Agency
  • Public
• Is the impact one-time or recurring?
• Is the impact quantifiable in financial or operational terms?
• Is the use primarily internal or external?
• Does this align with the governor’s priorities?

Measurability

• Is there a clear baseline to compare against (e.g. current process times, costs, error rates)?
• Are the metrics quantifiable and objective?

Scalability

• Can the use case be scaled across multiple departments, agencies or processes?
• Are there any dependencies or risks that could limit scalability?
• Is the use case replicable across similar processes or functions?