Chelo runs a trading psychology practice. Instagram brought him leads every day, and every one of them went through him personally. Here is the workflow we built, how it is wired, and an honest note on what we measured and what we did not.
Chelo consults traders on the psychology side of their work. One person, a personal brand, an audience on Instagram that kept growing.
The audience was the channel. A comment came in and Chelo replied. Someone asked for the lead magnet and Chelo sent the file. A prospect wanted a call and Chelo went back and forth about times in the DMs.
None of that is hard. It is just that it happened every day, for every person, and only he could do it.
That is Chelo's own count of the time going into DMs, comments, sending materials and coordinating calls. The three steps repeated in the same order every time: someone shows interest, someone wants the file, someone wants a slot.
There was also no filter. The same energy went to people who were nowhere near buying and to people who were ready, because both arrived in the same inbox.
“I was having a lot of stress because I was spending a lot of time and energy doing tasks manually - responding to DMs, comments, booking calls - and not getting the best qualified leads.”
Chelo · Trading Psychology Consultant
2–4 hours every day. DMs, comments and booking handled by hand.
Lead magnets sent one by one. Every request waited until Chelo got to it.
Leads recorded by hand, or not at all. Nothing connected the comment to a list anyone could work from.
Worth saying plainly, because the two get sold as the same thing.
Nothing in this system reads anything, interprets anything or decides anything. A person comments a specific word, and from there the steps run in a fixed order that we and Chelo defined up front: record the lead, send the file, offer a time.
That is the right tool for this particular job. The work we were removing was identical every single time, and identical work is exactly what a scripted workflow handles well. Putting a language model in the middle of it would have made the same three steps slower, more expensive and less predictable.
We do use AI where the work needs reading and judgement. In the RE Innovation project it goes through ownership records, debts and court documents for each property. There was nothing of that kind here, so there is no AI in this build.
A personal business should still feel personal, and a consultant whose product is his own judgement cannot hand the conversation to a bot.
So we separated the steps where Chelo makes a decision from the steps where he does not. Replying to a real question is a decision. Sending the same PDF to the two hundredth person who asked for it is not. Neither is copying a username into a list, or reading out available times.
Only the second group was automated.
Trigger · Instagram
Someone comments a specific keyword under a post. The Instagram API fires a webhook into Make.
Capture · Make to Airtable
Make creates a record for that person in Airtable: handle, which post and which keyword brought them, and when. No spreadsheet, no typing anything in later.
Deliver · the lead magnet
The material goes out as a direct message right away, whether Chelo is at his desk, on a call or asleep.
Book · Calendly
The same message carries a booking link. The prospect picks a slot from Chelo's real calendar instead of negotiating times in the DMs.
Land · back in Airtable
The booking returns to the same record, so the lead, the source and the call sit in one place instead of three.
The whole stack
Instagram → Make → Airtable → Calendly
Chelo is still on every call, and he still answers anything that is a real conversation. What runs without him is the part that was the same every time.
Case studies usually stop before this part. This one should not, because the honest answer is mixed.
What we can stand behind
Time. Chelo puts the manual work at two to four hours a day before, and close to nothing after. That is his own count, not a number we instrumented.
Delivery. The lead magnet goes out in seconds instead of whenever he next opened Instagram.
Lead capture. Every Instagram lead now reaches Airtable with the post and keyword it came from. Before, some were written down and some were lost.
Time to live. Two weeks from the audit to a working system.
What we cannot show you
We did not track conversion from comment to booked call, cost per lead, or revenue before and after. There was no baseline for any of it, and we did not put one in place at the start of the project.
“After working with them, they helped me a lot to automate all that process. I have now better leads, scaling faster, and I feel more relaxed because I don't spend all my energy on that. I put all my time and effort into doing tasks that add value to people who are already my customers.”
Chelo · Trading Psychology Consultant
The full conversation is in the video above.
Show us what happens from the first message to a booked call. We'll look at which steps still need you and which ones don't.
Let's talk →30 minutes. No pitch.
AI cut property review time by 90%
Read case arrow_forward Costa Cars · Automotive95% faster paperwork per car
Read case arrow_forward Car Amigo · AutomotiveYour whole fleet. One system.
Read case arrow_forward 3D Device · E-commerce60 managers. One sales process.
Read case arrow_forward Yolo Real Estate · Real Estate10× more leads. Same team.
Read case arrow_forward All Case StudiesSee every result across the team
View all arrow_forward