Why Starting With AI Economics Is a Costly Mistake

This week I’m going solo to talk about what I think is one of the biggest mistakes companies will make with AI: starting with the economics. When you look at what AI agents can do and start asking whether 20 people could become 10 or 5, you end up calculating the savings before you’ve redesigned the business to actually capture them. Using a customer success team of ten, I walk through what goes wrong when you cut to three, from lost context and stuck handoffs to one overwhelmed person holding everything together. Then I show a better path: start with the work, responsibility, execution, and infrastructure, and let the economics emerge from the design. I close with an exercise you can run on any AI calculation you’ve already done. If you’re an investor or operator weighing AI labor in your business, this one is for you. Episode Timeline & Highlights \[0:00] – The biggest mistake companies will make with AI: starting with the economics \[0:23] – Could 20 people become 10 or 5, or handle three times the volume, and why that math is tempting \[1:28] – A ten person customer success team and what AI agents can actually do around the clock \[2:24] – Before you cut to three, did you understand the work, the judgment, and the exceptions those people carry \[4:02] – The spreadsheet looks great, then handoffs break, context disappears, and exceptions pile up \[4:31] – The one remaining person who becomes the overwhelmed hub holding everything together \[5:28] – The hidden assumption: that your organization is just a collection of work done by people \[6:40] – Economics first on the growth side: can the business absorb ten times more opportunities \[7:38] – Why economic architecture sits in the AI labor architecture framework as an outcome, not a pillar \[8:50] – Start with the work, then responsibility, execution, and the infrastructure that connects them \[10:57] – Redesigning the ten person team with the economics taken off the table \[12:30] – Maybe three people can do it, or maybe ten people with AI agents can serve five times the customers \[14:17] – Beyond payroll: speed, consistency, capacity, availability, and where people create the most value \[15:21] – Breaking the stubborn link between growth and headcount, and what happens if the architecture is wrong \[19:06] – The exercise: set your AI math aside and go underneath it to the work, the people, and the infrastructure \[20:32] – Why economic architecture comes at the end of the framework and has to be earned Key Takeaways 1. Don’t Start With the Spreadsheet: Calculating AI savings before you redesign the business means your math rests on assumptions you haven’t tested. 2. People Carry Invisible Work: The people on your team hold context, make decisions, and move work between systems, so cutting them without understanding that work breaks the infrastructure. 3. Economics Is an Outcome: Economic architecture sits at the end of the AI labor architecture framework because better margins and capacity come from redesigning the work, not from declaring a number. 4. Redesign Before You Resize: When you understand the work first, you may find three people can do what ten did, or that ten people with AI agents can support five times the customers. 5. Break the Link Between Growth and Headcount: With the right architecture, growth no longer has to mean more people, more handoffs, and another layer of managers. Links & Resources - smrtPhone: https://www.smrtphone.io - That Real Estate Tech Guy: https://thatrealestatetechguy.com Pick one place in your business where you’ve already done the AI math, set the number aside, and go underneath it to the work, the people, and the infrastructure. You may find the real opportunity is bigger than the savings you first imagined. More high-signal conversations coming next.