Three ways I help

AI Readiness and Value Map

The question it answers: Where does AI actually earn money here, and where would it waste our time?

Four to six weeks. I map your workflows, identify where AI creates measurable value, and rank the opportunities by risk, feasibility, and payback. You get a sequenced plan, an estimated savings figure for each use case, and an honest list of what you should not attempt yet.

Best for: organizations with executive interest in AI and no agreed starting point.

AI Governance for Utilities

The question it answers: If a regulator asked how we govern AI, what would we hand them?

Almost every utility is now using these tools, and most have nothing written down. That gap is a live exposure, and state commissions have started asking.

I build the framework: use-case classification by risk, human-in-the-loop requirements by tier, documentation and validation standards, model inventory, and a defensible answer to the regulatory question before it's asked.

Best for: utilities with AI already in use and no formal policy.

AI Assurance and Security Review

The question it answers: Can we trust the AI system we just bought?

Independent evaluation of AI systems already in your stack — vendor-supplied or internal. Where they're reliable, where they degrade, what happens under conditions they weren't trained for, and how they could be manipulated.

This draws directly on my research background in adversarial attacks against power system models.

A note on independence: an auditor cannot also be a vendor. I don't sell you the system and then review it.

Best for: procurement decisions, post-deployment review, and any system whose output feeds an operational or financial decision.