Careful AI Use Leads to Profits

I’ve been pushing all my clients toward AI this year.

I’ve been pushing myself, too.

Not because it’s trendy.

Because for most service businesses, AI can do transactional knowledge work faster, cheaper, and at a higher quality than the person currently doing it.

It’s faster to “hire.” You’re not posting on Upwork, reviewing proposals, doing test projects. You can have an AI agent working on something within minutes.

It’s faster to train. You tell it what you want, give it feedback, and it adjusts. No onboarding. No ramp-up period.

It works on your schedule. Automated. Consistent. Available at 3am if that’s when you need it.

And it doesn’t just match human output. In a lot of cases, it exceeds it.

This isn’t just about cutting team members, though that is a real benefit.

What I’m talking about is the type of work you might hand to a virtual assistant, a cheap overseas freelancer, or a consultant.

Triaging your emails. Checking the status of everything in your project management tool. Pulling data together and putting it into a report.

That’s transactional knowledge work. And it’s being replaced across the board.

One of my clients is building an order planning assistant. It won’t place the orders on its own, but it does 80% of the prep work.

That means they can order more frequently. Twelve smaller orders a year instead of two big bets.

Then ship by sea instead of air, which cuts transport costs significantly.

Lower risk. Lower costs. Better margins.

Several of my marketing agency clients are reducing costs, increasing output quality, and hitting deadlines more consistently.

Their retention will improve. Their margins will improve.

If you can get the information from somewhere, and you can connect it to the thing it needs to output to, anything in the middle can be done.

It’s a genuinely exciting time.

The frenzy phase is real. Don’t skip it.

Here’s what I’ve noticed with almost every person I’ve introduced to AI, including myself.

There’s an initial frenzy.

You start playing with it. You get excited about the possibilities. You try a bunch of things, some practical, some silly, some low-impact but fascinating.

You’re attacking this new thing from every angle, testing what it can do, figuring out how it works.

This phase is essential. Don’t skip it.

It’s the best way to learn. You’re excited, you’re experimenting, and you’re building an intuition for what’s possible.

By the end of this phase, you’ll have figured out something important: this is a very smart, 24/7 worker that can work ridiculously fast, but it has no context and it needs you to give it some guidance.

That’s a powerful conclusion. Because once you understand that, you can start directing it properly.

Then you have to rein it in.

For most people, the frenzy lasts four to six weeks.

After that, one of two things happens if you don’t intervene.

Either the frenzy continues, and it turns into productive-feeling procrastination. You’re busy with AI, but nothing is actually moving the needle.

Or you crash. The excitement fades, the reality that AI can’t do everything sets in, and you go back to doing things the old way. You don’t realise any of the benefits.

Neither of those is good.

The way to prevent both is guardrails.

At some point, four to six weeks in at most, you start asking yourself three questions before working on something with AI:

1. Does this have a big ROI for me, my team, or my customers in the next one to three months?
2. Is this anywhere near the top of my priority list?
3. Is this procrastination disguised as productivity?

Simple test. If you can’t answer yes to the first two and no to the third, park it.

The arc I keep seeing

I’m watching this play out across multiple clients right now.

Frenzy. Excitement. Then either overwhelm, a crash, or the frenzy bleeding into unproductive busyness.

The real value starts when you move from experimentation to focused implementation.

Lock in the wins. Stop chasing the shiny stuff.

The frenzy is a great phase. If learning was this exciting in school, we’d all be in different positions.

I’m not saying avoid it.

I’m saying: go through it, and then put the guardrails in place so the excitement turns into actual profit.

If you’re in the frenzy right now, or you’ve crashed and given up on AI, I can help you figure out where the real profit is. Book a call and let’s talk about it.