Boards are approving AI they can't see.

Banks are putting AI agents to work faster than their boards can oversee them. These agents don't wait to be told what to do. They research, flag, draft, message, and recommend on their own.

Board oversight was built for people and for software that follows instructions. The gap is the distance between what the agents are actually doing and what the board can see, question, and control.

Closing it doesn't take a technologist on every board. It takes a director who has run AI in the field, knows where it fails, and knows which questions to ask before something reaches a customer or a regulator.

I run an AI-first sales team every day. My rule is to assume the agent got it wrong. I make it reread my intent and find its own misses, then make it validate every fact it gave me. Only then does a person spend time on the output.

That's board oversight in miniature: clear intent, a standard the system has to measure itself against, and human judgment where it counts.

• 13 years in lending and banking, so I understand the business model, the lending cycle, and the regulatory reality.
• 11 years advising financial institutions on CRM, data, and AI decisions, including how those projects go wrong.
• Five years leading a regional sales organization at Salesforce, accountable for an eight-figure annual number.
• Hands-on experience running AI agents in daily work, not just reading about them.

Where I fit

Community banks, credit unions, and financial technology companies. I'm most useful on technology, risk, and strategy committees, or as the director who helps the board ask better questions about AI.