Collections or diabetes: why I chose to build Remi

I spent four or five months choosing which company to build before starting anything. Eight rounds of filters: 112 ideas became 7, then 4, then 2. The full rounds are in From 112 ideas to 7.

It has a name now: Remi.

What Remi is

Remi is a monthly membership in Mexico for people with diabetes, high blood pressure, or weight issues.

The membership includes:

• An AI companion on WhatsApp.
• Remi doctors and specialists.
• Remi-brand medications, delivered to your home.
• Connected devices.

It all starts with a free at-home diagnosis.

How it came down to two

I went from 4 to 2 by modeling each candidate: the numbers and the projections, to see what each one would need to reach 1 million in EBITDA. The ones that looked unrealistic were out.

Two very different ones were left.

The other finalist: AI collections

It was a really hard decision.

The other one was a system that gives any SMB an AI-powered accounts receivable department. Over Slack or WhatsApp you say "send the invoice to Juan Pérez at this company" and it sends it on its own. Then it handles all the follow-up so you get paid, using the standards that reduce how much ends up in collections.

If an invoice goes more than 60 days past due, our collections department steps in, with an agency, to collect it.

The idea was to make it super easy for any SMB to have a full accounts receivable and collections department.

The number that tipped it: more than 2,000 customers

In the model, both looked very good.

The difference was how many customers each one needed. To get to the same place as Remi, the collections company needed more than 2,000 customers. And 2,000 B2B customers are much more expensive and harder to get.

That was one of the biggest reasons.

The other goes hand in hand: I think I can launch and grow faster with a B2C than with a B2B.

The problem

I like the problem. It can help a lot of people, and that matters to me.

It's a need with very strong pain. In English they call it a hair on fire problem: people need it now.

Mexico has some of the highest rates of diabetes deaths, diabetes-related amputations, and childhood obesity. There's a lot to help with there.

What worries me most

Being able to offer an affordable membership, within reach of a significant number of people, and making the numbers work.

According to the model, it should work.

What's next

1. The architecture: how the process and the company will work.
2. Tech and legal, at the same time: all the technology and all the legal work that needs to get done.
3. The real structure: taking it from paper to actually building it.

Every step will be on the Remi page and in the log. If you want to use the same method to choose a company, the template is in Resources.