I refuse to learn the same lesson twice.

Halfway through a call

I made my first cold call ever at the end of August. Benjamin and I had decided to learn the work by doing it, because we didn't want to get stuck learning before we started.

Two weeks in, I was on a call when I heard myself and thought, what the hell are you saying? I'd fumbled my sentence to a receptionist. She understood enough to put me through, and while she did, I realized I couldn't say the same thing to the next person. I didn't know why I was calling.

Years of restaurant work taught me to improvise, but on these calls improvising meant fumbling out loud with a stranger listening.

We weren't selling anything. We didn't have a product yet. I called to ask one thing. What manual work comes back every day? I was calling seven to eleven different industries. Before each call I did my own research and had an AI fill in what's common in that line of work. I'd kept feeding it more, until one company came with five different guesses about what was wrong there. Half of it I had to look up before I understood it. The research had turned into a mess. I was researching the research.

Before the next call I stopped. I changed what the research could bring me. Only what the company's own website says in black and white, which means what they make, how it works, and where it hurts. If it isn't on the site, it stays out.

The next call went fine.

That fix helped one person. Me.

Note to self

On September 10 a company I'd called rang me back. They had a system that worked for them. They thought we wanted to take it away. I logged the call and added one line. "Note to self. A lot of people take it as us trying to replace their system." The lesson was good. It had one reader.

Benjamin comes out of every day with something new too. He listens to an absurd amount of podcasts on sales, products, and running a company. I'm not there for most of them. When his view changes, it takes me a while to catch up, and I live in his house. Now picture a salesperson we hire from outside.

Where do the fixes at your company end up?

One fix for every employee

In Mellovy, an Employee is an AI agent that works in your tools on its own computer. When it gets something wrong, the fix doesn't end up in a note. You correct the Employee once. The fix becomes an Evaluation, and every Employee at your company answers to it.

The Evaluation doesn't take the Employee's word for it. It checks your systems to see whether the email went out or the order closed.

A lesson one person learns should be a lesson for everyone. Benjamin and I refuse to learn the same lesson twice. We're building Mellovy on what we've learned, for you as much as for us.