At The Agent Company, I’m turning a broad AI platform into an offer people will pay for. Working notes from day 45.
day 45 · CRM, lead source, outbound from nothing · 5 demos · 4 paid pilots
A working product, a serious engineering team, and almost no go-to-market.
That's the normal shape of a technical founding team, and it's the shape I'm most useful in.
Outreach ran through several pieces of stitched-together tooling with no system of record behind them. No one could answer "who have we talked to and what happened."
The platform could do many things; nobody had picked which one to sell first. Without a named problem and a named buyer, every conversation started from scratch and none of them compounded.
The product decision came out of the sales motion, not ahead of it.
Building the demo is how we found out what people would pay for. Four moves in 45 days, in the order they happened.
Local beat email on response rate, by a lot.
Five demos, four paid pilots, and what I'd tell you about them.
From a standing start, with no inbound and no brand. The demo-to-pilot rate is the number that matters at this stage. Pipeline volume isn't.
Same offer, same pages. Being in the room moved the response rate more than any copy change did. G2i's events channel had already taught me that.
Paid pilots, priced so the constraint is projects built, not deals closed. The price gets revisited once the pilots show which problem is worth a full one.
Pilot-to-paid conversion, retention, and whether the outbound agent earns its keep once the local channel saturates.
It's the job I'm interviewing for, done in public, with a date on it: diagnose the business, decide what to sell, build the operating layer, get in the room.
Everything above is as of day 45, September 2026.