SetDM

• 

  Dashboard

• 

  Autopilot is queueing messages for the user

• 

  Automated responses for conversions

Why We Built It

We kept seeing the same pattern. Fitness coaches were not struggling with traffic. They were drowning in unreplied DMs, cold followers, and missed buying signals.

We manually ran their inboxes with ManyChat and scripts. It worked. We generated $78,000 across clients. But it was predictable, repetitive work. If something is predictable, AI can do it better.

That was the spark.

What We Learned

Coaches rarely need more leads. They need a system that turns existing followers into booked calls.

Human setters are inconsistent. AI is consistent, scalable, and compoundable.

Niche training beats general training. A model trained only on fitness conversations outperforms everything else.

How We Built It

Collected thousands of fitness DM conversations

Broke them into patterns like qualification, goals, objections, and money talk

Built inbound and outbound frameworks from proven scripts

Trained a compounding AI brain only on fitness interactions

Structured everything in a setter agency workflow:

Lead→DM→Qualify→Nurture→Call

Challenges

Turning messy human DM logic into clean states

Making AI sound like a real coach instead of a bot

Mapping objection handling without breaking conversations

Balancing automation with human override when needed

Ensuring the model improves across all accounts instead of only one

The Outcome

A fitness specific DM AI that runs inbound, runs outbound, qualifies leads, follows up automatically, and books calls without human labor.

And it performs better than a human setter.

Built With

• nextjs
• python
• react