An honest story about how the Y Combinator principle works: "Do what doesn't scale"

I myself was the founder of 2 startups, one failed, the other successfully sold. And after this long journey, I'm wildly annoyed by business consultants, dumb advisers, and so on. Who don't understand how the real founder's score goes from 0 to 1. And I decided to kill the market of consultants and bloggers, the very ones who are telling you something.

The idea: "What if there was a startup consultant in your pocket?"

Reading Paul Graham's essay is cool, but when you have deadlines, you need fast, personalized answers.

Hypothesis, If you package Y Combinator's knowledge into an AI agent that is always available, it will be useful to other founders.

MVP: Do what doesn't scale

What's in the AI agent's head:

- All lectures from YC School (600+ minutes of video)

- An essay by Paul Graham

- Techniques from successful YC companies

- A database on the pivots of successful startups

- Specific frameworks for hypothesis testing

The principle of operation is to respond only based on verified sources. No general advice, just specifics with examples.

Launch: manual search for the first users

What did you try:

, Zen articles — 0 click - throughs

, Articles on VC.ru — 0 transitions, maybe something will change

✅ Entrepreneurs' chats in Telegram— it works!

What I did: I searched for relevant chats, shared the agent where it was appropriate

Twitter — just started

Strategy: A different audience, more international founders

Numbers after the first month

Audience growth:

- 29 subscriptions to the Telegram channel

- 100 unique conversations with the agent

- An average session of 8-12 questions in a row

Retention:

- 70% of users are returning

- They are actively used, 40% ask questions weekly

- Recommended, 3 users brought friends

Finance:

- Income 0₽

- 0₽ donations (not requested yet)

- Expenses: ~$40/month for OpenAI API

What I learned about users

Top 4 questions for the agent:

1. Give me an investor base?

2. B2B sales funnel

3. The structure of pitchdecks

4.the pnl template

What's next: I'm not giving up

Immediate plans:

1. Twitter development — there are more international founders

2. Adding new sources — materials from a16z, Sequoia

3. Monetization — possibly premium features for serious startups

What you need: I'm looking for contacts of entrepreneurs and chats with them!

If you know active communities of founders, please share them in the comments or in the BOS.

P.S. If you are a founder and want to test the agent, welcome! Write to me in the comments or BOS. And if you know the chat rooms of entrepreneurs, I will be very grateful for the contacts.

Follow the principles of Y Combinator: build, test, don't give up!🚀

Do you want to try ScaleMate???

tg project channel https://t.me/invmatch

AI agent https://dan2004005 .github.io/adviser-/ai%20adviser%202.html