Use cases are clustered into solution categories. Each dot on the matrices represents one solution category — sized by the number of backing use cases, colored by dominant industry. Matrix 1 (Readiness vs. Impact) identifies quick wins. Matrix 2 (Evidence vs. Fit) separates substance from hype.

Ranked solutions

AI score ranks each use case against our "How might we" goal — using LLMs / agentic AI within existing systems to cut customer & staff friction at a conservatively-run, regulated utility. Composite = 0.40·CX + 0.35·Fit + 0.25·Ev (each 1–5): CX = friction / customer-experience value, Fit = adoptability for our utility, Ev = evidence strength.
Verifiability is a separate check (top use cases web-verified): ✓ Verified = corroborated by a citable source (use the “cite” link); ? Needs research = real but only vendor-sourced; ✗ Unverifiable. Most sources here are vendor/marketing — triage, don't cite blindly.