Prompts are dead. Intent is the skill.

Building an account plan used to take me hours. I'd pull what the bank already owned, map every stakeholder and whether they were for or against us, then read investor filings, shareholder reports, local news, and podcasts to find what they actually needed. Then I'd handwrite personal notes to each person and build my own follow-up cadence.

And that was for one product. To sell a second product into the same bank, I'd do it all over again.

Today I hand an AI agent my intent: sell my team's products, have a point of view, build an outbound campaign to reach these people. It does the research like a PhD-level assistant, not a junior coordinator taking notes, and it covers every product at once.

The salesperson's job moved up. The agent does the labor. The human brings intent, point of view, and judgment.

No matter what the AI hands back, my first response is the same:

"Whoa, this is way off. Go back and read my original instructions."

I don't tell it what's wrong. That forces it to measure its work against my intent and find its own misses. When it comes back confident, I push again: some of this data doesn't look true, validate it. Only after it has checked itself twice do I sit down and read.

It's not about keeping a human in the loop on every line. It's about making the AI check itself before a human spends time on it. Students learn to do the same.

A hands-on, 3-credit course on how complex deals are won when both the buyer and the seller use AI. Students will:

• Direct an AI research agent with intent to build account intelligence that used to take a day by hand.
• Build use cases and outbound campaigns across several product lines at once.
• Run the two-pass check so the AI's work can be trusted.
• Practice the human part that still decides the deal: discovery, point of view, and the conversation in the room.

The final project is a live deal simulation against a regional bank buying committee. I can also teach a finance elective, AI and Fintech in Community Banking, drawing on my 13 years in the industry.

I'm open to teaching one or two evening courses per semester, in person or online. If a full seat isn't open yet, I'm glad to start with a guest lecture or a short series on the three levels above, and build from there.