My biggest takeaway from Endless Customers Live this week: when AI recommends a company, it starts with a pool of 30-40 options, then starts eliminating. Here's how it cuts the list, and the fix for each step 👇

Yesterday I spoke to a marketing class at UConn. Three student groups. Three real marketing projects. One of them is working for the university itself. The dean gave that team a big challenge: position the university as the best for experiential learning. That's what they want to be known for in the coming years. I love that goal. With how much the world is shifting right now, it's exactly where a university should be focusing. As they planned their research, here was my advice: Go to ChatGPT and ask, "What would a university have to do for you to recommend it as the best for experiential learning?" My guess is nobody at the university has asked that yet. What comes back will likely fall into two buckets: • Tactical moves. What to put on the website. Which student experiences and outcomes to document and share. • Fundamental shifts. What would we have to change? Where would we invest? What would we stop doing so we could go all in? That second bucket is the hard part. Being the best at something means making choices, including deciding what other things you'll pull energy, focus, and resources away from. I told the students to treat that first answer as a starting point. Keep probing. Stay curious. Shape it around what they heard from the dean. I'd bet they walk back in with something the dean hasn't seen before. Not a list of marketing ideas, but a clear picture of what it would actually take to own that position, from the "buyer's" point of view. And this applies to every business. Pick the one thing you want to be known for. Be specific about who you want to serve. Then ask AI: "What would a company have to do for you to recommend it as the best at this for these customers?" Then ask what evidence it would need to see. Then ask what you'd have to stop doing. Some of the answers will be wrong, incomplete, or unrealistic. Building a great business still takes human judgment. But you'll have a starting point you can question, test, and bring to your team. And you'll know exactly what AI says you'd need to do before it recommends you. Has anyone been doing this in their business? What did you learn, and what did you change?

They Ask, You Answer / Endless Customers didn’t work for my company until ______________. How would you answer that?

Bold prediction: within six months, ChatGPT and Claude will both launch a CMS. ChatGPT Sites already hosts websites on your own domain. Claude can publish live pages. Next: managing your whole site from a chat. An interesting alternative to WordPress or HubSpot.

Is your homepage the most important page for AI recommendations? I asked Gemini. Its estimate: • Service pages: 45-50% • Homepage: 5-10% Fix your service pages first.

The simplest AI visibility test I know, from IMPACT coach Mandy York: Ask ChatGPT to recommend the best company for what you do. If you're not there, ask why. Then ask what you'd need to do so it would. AI will literally tell you.

A buyer can understand your company, appreciate your content, and trust your people... and still ask: "What gets enough better for me to justify the effort and risk of changing?" That question is why "a business worth choosing" is Factor One.

One thing you can do this week to get recommended by AI more often: Write a fit statement. Who you're the best fit for, and who you're NOT a fit for. In our AI Visibility Study, fit carried more weight than anything else we looked at.

I almost let AI send emails on my behalf. Then Tom Nassr asked me: "Have you ever heard of this thing called prompt injection?" I turned the sending off. If AI can touch your inbox or accounts, start read-only. Let it draft. Keep the send button for yourself.