Challenge

Arizona Public Services was executing an AI assurance initiative across three use cases: regulatory intelligent search, energy theft analytics, and a personal admin assistant. Testing was largely manual, and the client needed recommendations for a test harness spanning data verification, model testing, security, integration, privacy, bias, fairness, and user experience testing.

Solution

QualityAI conducted a structured AI tool rationalization assessment evaluating more than 50 tools across three AI solution types against 86 test types defined in the client’s AI assurance guide. The team used an eight-dimension weighted scoring framework and stakeholder interviews to assess platform fit, cost, adoption, and governance constraints.

Impact

The client received a three-tier MVP toolkit recommendation with tool-to-test-type mappings, gap analysis, and license governance flags. Thirty-seven tools were recommended across RAG, predictive ML, and agentic AI, covering 100% of applicable test types with no critical gaps.