Every small business I've ever worked with has the same instinct when they hear "free." What's the catch? It's a healthy question. Here's the honest answer.
I build the first AI agent for you at zero cost. No deposit, no proposal fee, no discovery retainer, no "we'll waive it if you sign." Actually free. If it doesn't move your numbers, you owe nothing and we shake hands. If it does, you go onto a monthly retainer. That's the whole model.
Most consultants would never do this. Most can't. Understanding why not is the whole story.
The real cost of a bad AI project
When a small business buys a typical AI engagement, they're not really paying for software. They're paying for risk. The agency wants $15,000–$50,000 upfront because, honestly, they don't know yet whether what you're asking for is going to work. The upfront is how they get paid even if it doesn't.
You, meanwhile, are the one taking all the risk. You've handed over cash for something that might not land. If it doesn't work, you got an invoice and a Slack archive. If it does work, brilliant — but now you've spent the money anyway, and you only find out six months in whether the ROI shows up.
There's a better model. Move the risk from you to me.
The commercial logic (yes, it actually works)
Here's the maths from my side. I spend roughly two weeks of work on a first build. My cost to deliver is my own time. If I'm right about the leak being real and the fix being valuable, the retainer that follows pays me back many times over. If I'm wrong, I've lost two weeks and learned something.
The bet I'm making is this: I'm confident enough in my ability to identify a real revenue leak and build something that plugs it that I'll stake the time. If you don't see the numbers move, I wouldn't want you on a retainer anyway. Monthly retainers only survive if both sides are happy. A free first build is a filter — for both of us.
Big consultancies can't run this model because their overhead demands they bill every hour. I'm a solo operator. My overhead is a laptop and coffee. I can absorb a first build as the marketing cost of acquiring a long-term client. It's cheaper than running ads.
I bet my time. You bet nothing. If it works, we both win. If it doesn't, only I lose.
Tell me where you're leaking money →Why this is actually the cheapest way to try AI
Think of it like this. You can't know in advance whether an AI agent will actually recover, say, those after-hours phone calls you're missing. You can model it, estimate it, eyeball it. But you can't know until the thing is live and running for two weeks.
Under a traditional paid engagement, you pay to find out. Under this one, you find out first, then pay if it worked. It's the same information, but the sequence is reversed — and the sequence is the whole thing.
Another way to look at it: what's the most expensive way to try AI? Paying for it to fail. This model removes that path entirely.
What the catch actually is
I won't pretend there's zero catch. There are two, and they're both real:
- I have to believe it'll work. Before I agree to build, I'll spend an hour or two understanding your specific leak. If I think the probability of moving your numbers is low — because the leak isn't AI-shaped, or the data isn't there, or the scope is wrong — I'll tell you no. I'd rather pass on the build than waste both our time on something that won't land. If your situation is one I know well, I'll say yes quickly. If it's novel, I might say let's do a paid scoping week first.
- You have to be ready to act fast. A 14-day build only works if I can get you on the phone, see how you work, and iterate without a week of silence between messages. If your business isn't ready for a short, focused burst of attention, this model will break. Most clients are fine; a few aren't.
That's the entire catch. No "free trial that becomes a contract." No surprise invoices. No "only the first hour is free."
What happens after the free build
Two weeks in, the agent is live. You've seen it handle real calls, real leads, real quotes. At that point we have a short conversation about numbers. If it's moved the needle — and for most businesses it has — you go onto a flat monthly retainer that covers maintenance, improvements, and the next build. If it hasn't, we stop there.
I've covered the day-by-day of the build process in a separate post, and what the retainer actually includes in the FAQ. Both are worth a read if you're closer to hiring than not.
The bottom line
Most consultants sell you their confidence. I'd rather prove it. The free first build is how I put my time where my mouth is — and how you find out, for nothing, whether AI actually does anything useful for your specific business.
If that sounds like the shape of deal you'd take, the next step is trivial. Send me one sentence about where your business is losing the most money. You'll hear back inside a day.