How I Built An AI Support Agent For Founders Making $500/Month [Side Hustle]
Who are you and what business did you start?
Hi, I'm Andy, one of the three co-founders of HeyDeacon in London. We make Deacon, an AI customer support agent for founders and small software teams who still answer support themselves.
Deacon answers customers on your website from your own docs and links the page it used. When your docs don't cover a question, it says so, takes the customer's email and hands you the conversation instead of guessing. Every question it couldn't answer goes on one list, and an answer you write once is used from then on.
We started it as a side project in June 2026, and today it brings in around $500 a month.
How do you come up with the idea for HeyDeacon?
The idea came from a problem my two co-founders already had. They run a couple of other software products, and both were getting more customer messages than they could keep up with. A lot of them were simple, repeated questions.
They figured we could build an agent to handle the simple FAQs first and then grow from there. We built Deacon as a harness around the language model: it finds the answer in your docs, says when it doesn't know and hands the conversation to a person. As the models get better, we can slot the new ones in and Deacon improves with them.
I've been a software engineer for the last 10 years, at several companies of different sizes, with a lot of building and tinkering along the way.
We didn't have to validate the idea in the usual way, because two founders already needed it. We built it for them first, with the goal that other like-minded founders would need it too.
How did you launch HeyDeacon and get initial traction?
We haven't done a big public launch yet. Deacon was built for my co-founders' products first, and since then we've been telling people about it directly. A Product Hunt launch is on the cards.
The reaction so far has been encouraging. Most people agree with the problem straight away, ask lots of questions and are impressed when they see it working.
Our first paying customer didn't come from any of that. They found us through organic search: someone in Japan who runs a small SaaS and needed help handling their customer requests. That showed us early on that good SEO pages can bring in paying customers on their own.
What was the growth strategy for HeyDeacon and how did you scale?
Growth so far has come from three things: outreach, search and showing the product working.
First, lots of outreach. We talk to founders on Reddit, Hacker News, Indie Hackers, LinkedIn and X. The people who need Deacon most are founders who answer support themselves, and those are the places they already spend time.
For search, we've put a lot of work into quality SEO pages. There are comparison pages for the support tools people already know, pages for the platforms founders build on, like Lovable, Bolt and v0, and guides that answer the questions founders actually search for. Our first paying customer found us that way. We've also listed Deacon on lots of free software directories, which helps new people find us and helps our own pages rank.
The third piece is showing rather than telling. A support agent is hard to judge from a list of features, so we make videos and images that show Deacon working: answering from a company's docs, linking the page it used and handing over when it doesn't know. Seeing it work impresses people far more than reading about it.
My advice is to pick the few places your customers already spend time and keep showing up there. Then back it up with pages that answer what people are searching for, because good pages keep bringing people in long after you've written them.
What were the biggest lessons learned from building HeyDeacon?
Honestly, my biggest mistake was not starting earlier. I spent 10 years as a software engineer at other companies, and I wish I'd started something of my own sooner.
My top tip is to build for someone who needs it right now. Deacon was built for my co-founders before anyone else, so we never had to wonder whether the problem was real. If you know founders with a painful problem, start there.
The other lesson is to build the part that lasts. AI models keep getting better, so we built Deacon as a harness around them: finding the right answer, saying when it doesn't know and handing over to a person. Each better model makes Deacon better without us starting again.
Discover Similar Business Ideas Like HeyDeacon
More about HeyDeacon:
Who is the owner of HeyDeacon?
Andy is the founder of HeyDeacon.
When did Andy start HeyDeacon?
2026
How much money has Andy made from HeyDeacon?
Andy started the business in 2026, and currently makes an average of $6K/year.
More Case Studies Like This
Starter Story × HubSpot
Free tools for every part of your business.
Find customers, win deals, and automate the busywork with HubSpot's free tools.
