The Truth About AI Consulting (After 87 Clients and 4 Years)

Over the last 4 years I've spoken with thousands of business owners about AI and automation.

I've implemented it for 87 of them.

My company got featured at the New York Stock Exchange and I've spoken on an AI panel about what's actually working inside businesses right now.

I'm not saying that to flex, but because everything in this article comes from doing the work, not watching other people do it.

And if there's one thing 87 clients taught me, it's this: being successful with AI comes down to having a good process. You don't need to be technical at this point.

I've made every mistake there is. My first project was $500. Then I charged $45 an hour. Then $60. Then $75. Eventually I figured out milestone projects, then maintenance and dev retainers. And even then I kept screwing up. I once converted a perfectly good contract into a dev retainer and lost the client, because a monthly payment with nothing visible behind it feels like a bill.

Every one of those mistakes taught me something that's now baked into how we run 11 active implementations at once.

If you're building an AI consulting business right now, or thinking about starting one, this is the 4-year shortcut.

First, let's clear something up

AI consulting. AI Services. AI agency. AI automation agency. AI consultancy. AAA. People argue about these labels like they're different business models.

They're not. It's all the same job: you get paid to make a business run better using AI and automation.

Some weeks that looks like consulting, mapping how a company actually operates and showing them where the leverage is. Some weeks it looks like an agency, building systems and shipping deliverables. Most of the time it's both, because the map is worthless without the build and the build fails without the map.

The label doesn't matter. What matters is that under every version of this business, the client is buying the same thing: a company that operates fundamentally better after you than it did before you.

Keep that definition in your head. It's the measuring stick for everything else in this article, and it's exactly what the people I'm about to describe don't deliver.

Why a lot of people are going to get burned at this

The AI agency space is full of people teaching a business model they've never actually run.

You've seen the posts. Sign clients with no experience. Deliver with no skills. $10K a month from a laptop, and the whole thing runs itself.

Here's the filter that will save you months of wasted time: if someone is showing you how much money they make and the whole thing sounds effortless, it's almost never adding real value to the client. And offers that don't add real value don't last. The client stays a month, maybe two, realizes nothing changed in their business, and cancels.

Anything that sounds effortless in this business is a lie you're being sold.

Real implementations are work. You have to understand how the client's business actually operates, which is usually different from how they think it operates. You have to map processes, sit in the mess, and build things that people will actually use. Nobody selling a shortcut talks about adoption, because their clients never stick around long enough for adoption to matter.

And this is how people get burned twice. First you pay for the course teaching the effortless version. Then you go run that playbook on a real business, the client cancels in month two, and you're back at zero with a burned reference in your market.

The clients get burned too. Every failed AI project makes that business owner harder to sell for the next person who shows up, even the ones doing it right. The larpers are salting the ground for everybody.

So before we get into the right way, get this straight: the money in this business is real. I've built a life on it. But it comes from fundamentally changing how companies operate, and there's nothing effortless about that.

The right way to do this

When you do this right, you fundamentally change how a company operates.

That's the frame most people never get to. You're not selling automations or chatbots. Whether the result is new revenue or hundreds of hours back, your impact goes to the foundation of the business. That's why clients pay real money for it, and that's why they stay.

Here's a live example. One of our current builds is for a 15-person home building company. The operating system alone is going to double their business. Not the AI. The operating system. It removes all the guesswork from their day-to-day and streamlines the entire company into one place. Every job, every schedule, every dollar, visible and connected.

Once that's live, we layer AI and automation on top of it.

That order is the whole method:

Step 1: Get everything in one place. One source of truth for the business. Most companies are running on 15 disconnected tools, spreadsheets, and someone's memory. Until that's fixed, AI has nothing solid to stand on.

Step 2: Layer agents and automation on top of clean infrastructure. Now the AI actually works, because it's plugged into real data and real processes instead of chaos.

Most people run this backwards. They bolt AI onto a mess, the outputs are wrong, the team stops trusting it, and the project dies. Then everyone blames the technology. The technology was never the problem. The order was.

One more thing that saves you from over-engineering: you don't always need to build the foundation. If a client already has a solid ERP or real infrastructure, don't rebuild anything. Build the agents and automation on top of what they already run. Their team keeps working where they've always worked, and the manual parts of the job start disappearing underneath them.

Sometimes you're selling the full operating system. Sometimes you're selling just the intelligence layer on top. What you're always selling is the same thing: a business that runs better after you than it did before you.

Do that, and retention takes care of itself.

There's no best way to get your first client

This is where most people stall, because they're looking for the one correct answer. There isn't one. There are a handful of proven paths, and the right one depends on your situation and what you're good at.

Warm network. Make a list of every business owner and operator you know. Reach out to all of them. You're not pitching, you're asking what's eating their time. Book calls from there. If you have any network at all, your first client is probably already in your phone. This is the fastest path for most people.

Upwork. No network? Set aside $300 to $500 for connects, optimize your profile, and send proposals every single day. Do that consistently and you will land at least one client. It's how I started. Not glamorous, but you get paid reps fast, and reps are worth more than money at this stage.

Dream 100 outreach. Pick an ICP, build a list of 100 companies that fit, and create an offer specifically for them. Lead with a strong guarantee or do the first build cheap or free in exchange for the case study. Slower to start, but you're building a pipeline you control from day one.

In-person AI workshops. Massively underrated. Run a free or low-cost workshop for local business owners on what AI can actually do in their operations. You walk out of a single room with more qualified conversations than a month of cold DMs. Business owners buy from the person standing in front of them.

Cold outreach and content. Both work. Both take longer. Content compounds over months, cold email needs infrastructure and volume. Better as a second channel than a first one for most people.

What you're actually selling (this matters most on cold outreach)

The channels above split into two groups. Warm network, Upwork, and in-person workshops come with built-in trust. The person already knows you, the platform vouches for you, or you're standing in front of them. You can be a generalist there and still close.

Cold outreach has none of that. A stranger reading a cold email won't buy "we help businesses with AI." It's too vague to trust and too big to say yes to.

So if you're going cold, and you don't have a big personal brand doing the trust-building for you, you need an offer built around one specific thing you do for one specific type of client. One workflow, one system, one result. Something they can picture in their business before they ever get on a call with you.

That one thing is the hook, and here's the part people miss: it's the entry point, never the ceiling. You get in the door, you deliver that one thing well, and now you're not a cold email anymore. You're the person who already made something in their business work. That's when you expand into the full implementation, the consulting, the retainer.

Show them one thing that works. The rest of the relationship gets sold by the result, not by you.

Pick the channel that matches your situation. Then commit to it for 90 days. The people who fail here don't fail because they picked the wrong channel. They fail because they switched channels every two weeks.

And here's why your choice matters more than you think: how you get your first few clients shapes what you scale. Which brings me to the part almost nobody talks about.

Scaling past your first few clients

Everyone asks about scalability. Here's how it actually works.

Your first few clients are research, not just revenue.

Once you've delivered for 3 to 5 clients, look at what you have. Which project produced the clearest ROI story? Which vertical had the most repeatable problem? Which client was easiest to serve? Your best case study picks your vertical for you. You don't choose a niche on day one. You earn one.

Then you build the branch.

Take that winning case study and turn it into a machine. One offer, one ICP, outbound campaigns built around that exact result for that exact type of company. "We built this for a home builder like you and it's on track to double their business" closes 10x easier than any generic AI pitch, because the prospect can see themselves in it.

Depending on what your warm network looks like, that network might even decide which branch you start with. If half your contacts are in construction, that's not an accident. That's your first vertical.

And turn your clients into a sales channel.

You need a real referral program for this business model. Here's the one that works: offer 20 to 30% of the lifetime value of any client someone refers to you, paid ongoing as the revenue comes in.

That sounds expensive. That's the point. Early on, a new client is worth far more to you than the margin, because every client is a case study, a testimonial, and a source of more referrals. You're deliberately overpaying to maximize conversions when you need momentum most. And it costs you nothing until the revenue actually exists.

Once you're at around 10 clients, you don't need an offer that generous anymore. You'll have case studies, inbound, and a vertical machine running. Dial the referral percentage back and let the other engines carry it.

Scale in this business doesn't come from adding more channels. It comes from going deeper on the one that's already working.

The rules I'd give myself starting over

If I could hand my year-one self a single page, this would be on it:

Nothing in this business is effortless. If an offer sounds effortless, the value isn't real and the client won't last.

Process beats technical skill. Map how the business actually works before you build anything.

Get everything in one place first. Layer AI on top of clean infrastructure, never on top of chaos.

If they already have solid infrastructure, don't rebuild it. Build on top of it.

Price off the cost of the problem, never off your hours. I went from $500 projects to $45 an hour to milestone deals worth 50x more for the same skills.

A retainer needs visible value every month. If the client can't see what the payment buys, it's a bill, and bills get cancelled.

Your second client should look like your first. Case studies pick your vertical, then you go deep.

Overpay for referrals early. 20 to 30% of LTV until you hit 10 clients, then dial it back.

Save this. You'll need different lines from it at different stages.

Where this can take you

Four years ago I had a failed SMMA and a failed cold email SaaS behind me. I was sending Upwork proposals every day and driving Uber in Miami to cover rent.

Today Boom Automations has served 87 clients, got featured at the New York Stock Exchange, and runs 11 implementations at a time with a team of forward deployed engineers.

Nothing about that path required me to be a genius or a coder. It required doing this the right way when almost everyone around me was selling the easy version.

And here's what gurus won't tell you: the opportunity is bigger now than when I started. AI is more capable, businesses are more aware they need it, and the market is still flooded with people doing it wrong. Every larper burning clients is creating demand for someone who does it right.

That someone can be you. But only if you skip the shortcuts.

One last thing. Everything in this article is easier with people around you who are already doing it. I run a private group of agency owners building this exact business right now. Real builds, real sales calls, real numbers, shared every week. No theory, no larpers. You bring your offer, your pipeline, and your problems, and you get eyes on them from operators who are a few steps ahead of you.

If you want in, DM me "INNER CIRCLE" on X and I'll send you the details.

The playbook's in front of you. Go run it.