How I think about AI
Every AI project sits on the same graph: how hard it is against how much it changes the business. Here is how I decide what to do first.
I was discussing AI with a business owner friend of mine when he said: “I hate paying someone a hundred grand to do a $20-an-hour task.” It’s a great line. For many business owners, it’s even worse: WE’RE the ones doing that $20-an-hour task.
When it comes to implementing AI, after you get excited about how it can help you personally, then you start getting excited about how much money AI can save you. Which is great!
But run the math all the way out. Say you run an $8M business at a 25% net margin, and a sweeping AI effort lifts you to 40% net, which is the rosiest case anyone would dare put on a slide. Congrats! You just found $1.2M a year. That’s real, serious money and you should take it. But now try to squeeze out an additional $1.2M again next year from the same business. You can’t.
Cost-cutting runs until it hits a wall, and the wall is the business itself: same customers, same offer, same revenue engine. There are only so many costs to take out of a company to create profit. Past a certain point, you’re not improving the business anymore, and you may even start harming it. The only way to create additional profit past this point is new revenue. And for me, that is the ultimate goal of AI adoption.
Every AI project sits on a graph
When I think about an AI implementation, I ask two questions to get oriented: how hard is it going to be to execute, and how much does it change what the business is capable of? Plot those coordinates and you get the Leverage Graph, which I break down into three zones.
- Personal Fluency (bottom-left): the owner gets AI working in their own day. This part is easy, and sometimes it dramatically changes that person’s week. But it doesn’t change the company.
- Operational Efficiency (the middle): AI does what the business already does, but cheaper and faster and with fewer errors. The $20-an-hour math lives here. It changes how the company runs, and because this is known work, the ROI is typically very quantifiable.
- Revenue Growth (top-right): AI enables the business to do something it couldn’t do before. Bid eight jobs a week instead of one. Settle cases faster to increase case velocity. Get in front of twice the prospects you used to. Even: open a new line of business that didn’t exist last year.
Scroll the figure sideways for the rest of it.
Efficiency moves are capped by your expenses. Capacity moves are capped by your TAM.
That asymmetry is the point of the whole graph. The best margin year you will ever have is bounded by the business you already built. You can’t save your way past your own expense structure, and the returns diminish as you get more efficient.
But a capacity move’s ceiling is whatever the market will give you. If you optimize that $8M business hard: it hands you maybe $1.2M more. But if you grow that business to $20M and don’t even improve margin, it hands you $3M more, AND you can keep growing next year.
Scroll the figure sideways for the rest of it.
There’s a second asymmetry, and it’s coming whether you like it or not. Efficiency has a shelf life because everyone else will eventually adopt similar AI tools. Any tool you can buy off the shelf, your competitor can order next quarter, and the savings get competed away. But the new work you won while they were still optimizing for efficiency stays won. The jobs, the clients, the reputation, the data: those you keep, and continue growing with. So the long term advantage is not in saving money with flashy tools; the advantage is using those tools to GROW.
Scale numbers are scenarios. Efficiency numbers are arithmetic.
The payroll math is checkable. Since you know how much you pay a team member and how much time they spend doing a job, you can calculate fairly accurately how much ROI an AI tool to increase their efficiency will give you. A growth number is a scenario: it depends on win rates, demand, crews, working capital, and your follow-through. Both numbers belong on the table, but let’s not pretend that one of them isn’t speculative.
So when I help business owners think about how to invest in AI, here’s the simple way I attack it:
- I look for efficiency gains AI can provide where the savings can pay for the cost of implementation. If I lead you through a $20,000 implementation that nets you $100,000 in savings over the next year, most business owners I know are going to do that all day. And that part is largely predictable.
- Once the project is paid for, I look for opportunities to scale. This part is a bit more of a gamble; since you’ve never done it before, you don’t have real numbers to calculate an ROI. The best you can do is take an educated guess. But if those scale moves were essentially paid for with savings, then it is almost like it’s risk free. In the worst case, you end up with a financial wash, but you have moved your business forward into the future technologically. Best case: you’ve overhauled your growth engine for “free.”
In other words, savings pay for AI projects, but growth is the real reason to do it.
Don’t forget about the X-axis
Scale moves exist farthest right on the Leverage Graph because typically they’re also the most difficult to implement. It takes some conviction and excitement to push through a challenging AI implementation, even if you feel pretty confident about the payoff. And that’s why Fluency comes first. An owner who has personally felt what AI does for them is the one who recognizes the Efficiency move when it shows up. Personal leverage is the on-ramp, not the destination.
Margin moves come next. They’re fast, they’re provable, they train your team to work with AI, and the savings can often pay for the climb.
And THEN the Scale move, once you’ve experienced the exhilaration of watching AI do a job that used to take you 8 hours in 5 minutes.
Sometimes the assessment says skip the line and go straight at a Scale project. And if that happens, that’s awesome! In one of my pilot projects when launching Dominion AI, the first and most logical AI project was a hard top-right Scale system. At the time I’m writing this it’s still playing out, but most likely, this $30,000 project will create $1,000,000 in new revenue over the next year that would not have been captured before.
But when things progress more traditionally, the Leverage Graph is how a fifty-person company reaches the top right without betting the company.
Scroll the figure sideways for the rest of it.
When you do AI right, you move your bottleneck to a place you choose
The owners I work with all start with the same problem: a hundred possibilities with AI, they can’t find the first five hours to get competent, and they have a nagging feeling that the meter is running. What I’m an expert at is getting owners over that hump.
But once you get moving and grasp what AI can do in your business, you’re going to start creating new problems (the good kind).
For the company I built that pilot project with, the new system will soon mean they’re winning more bids than their existing crews can complete in a timely manner. Now they have a new bottleneck at the fulfillment layer.
If you actually harness the power that current AI systems can offer you, what you’re going to find is that your new job is playing whack-a-mole with bottlenecks where AI creates more leads or more sales calls or more new clients than you can handle. That’s an enviable problem to have, and solving it becomes your next AI project!
So you can summarize my approach to AI as: the savings pay for it; the growth is the point.
James Green, Dominion AI