Flip it around and it's obvious. If that same proposal had come with a guarantee attached, sign here and this puts back double what it takes out, you'd have signed it in the truck on the way home. The number was never the problem. Not knowing was.

I want to tell you what that unsigned proposal is actually worth now, because it changed recently, and almost nobody in the trades has noticed.

I spent thirty years in enterprise sales before I started building AI systems for businesses, including my own home repair company. A few weeks ago, I ran a test. I wrote the kind of scope of work a trades consultant sends a qualified prospect, the private version, not the vague stuff on their website. I'll be straight about how I built it: it's a composite rather than a copy of any one outfit's document, priced against published rates for that kind of work, because a scope of work without a dollar figure next to it is just a description. A 90-day receivables and invoicing fix for a plumbing company, about $9,500. New payment terms, a three-stage follow-up cadence at 5, 20, and 35 days past due, a card-on-file request at estimate signing. Then I gave that outline to an AI with one sentence of context: an eight-technician shop carrying $180,000 in receivables over 60 days past due.

In about a minute it wrote the invoice language, the day-20 payment reminder, the day-35 phone script for the office, and the card-on-file ask worded so it doesn't feel like a credit check. Not advice. The actual scripts, ready to use that afternoon.

Here's the part that matters more. Connected to the systems you already run, your field service platform and your payments setup, the way the newest one-window AI tools allow, it ends with a question no consultant's PDF ever asked: want me to load these terms and the follow-up cadence in now, or schedule it for the start of next month?

I call this the Kerzie effect: once you have a seller's scope of work and your own business context, an AI can synthesize, and now execute, what used to require paying for the seller's time. The proposal is the instruction set. You have the context. The AI does the work.

This is not a knock on coaches. The good ones know things no outline contains: which of your techs is coachable, what your market will actually pay, what breaks when you grow past ten trucks. That's observation, not a playbook: things a person standing in your shop can see that no document can. Accountability too, somebody who will come back and ask whether it worked. That survives. What doesn't survive is paying playbook prices for playbook work. Harvard Business School ran a field experiment with 758 consultants at BCG, one of the most expensive firms on earth, and found AI did their within-reach work about 25% faster at roughly 40% higher quality. If it works on their playbooks, it works on the ones being sold to you.

Two honest cautions, because I run a trades business myself.

First, a word about that competency line, because most owners picture it in the wrong place. Think of AI competency as a continuum. On the far left is the guy who says he's using AI, and what he means is he types a question into a chatbot and reads what comes back. On the far right is a fully AI-operated company, autonomous systems running the whole thing. The line you have to cross to do what I just described sits much further left than most people assume. It is past the chatbot, and it is nowhere close to the far end. It is also moving toward you, because the tools keep getting simpler. The line is real, though, and it cuts both ways: that same study found people working beyond the AI's reach did worse than people using none at all, by 19 points.

Second, crossing that line is not the only requirement. Somebody has to care. Care that $180,000 should not be sitting in receivables past sixty days. Care enough to say, I want to own this, and I want to make our operations better. A follow-up cadence nobody runs is a document, not a change. If nobody in your shop wanted collections to get better, the AI didn't fail you. And whatever you do build, you now maintain: cancel the consultant and the follow-through lands on you.

Here's the part I'd pay attention to if I owned a shop, because it isn't about the software at all. This is not job-title dependent. In a two-person photography studio I know well, the person who crossed the line was the owner, because in a shop that small there isn't anybody else. In a metal fabrication company with a real payroll, it was the controller, who saw it before anyone above him did. Neither one is technical.

What they have in common is a sentence I hear constantly now: I know we should be using AI in this business, and I don't know how. That's the whole qualification. It's somebody who has correctly worked out that their own knowledge is the only thing in the way, and who knows in their gut that if that changed, the business would run better. In your shop it might be you. Just as likely it's whoever already knows where the money is buried: the office manager who could recite your aging report from memory, the dispatcher who knows which jobs come back. They already carry the context the AI needs, and they don't need your permission to go learn the rest.

That cuts a way most owners haven't thought about. Almost nobody in any company crosses this line, which is exactly why the one who does becomes very hard to replace. I don't think the number of jobs in the trades is going anywhere; my read is that the titles stay and the skill underneath them turns over. If somebody on your payroll has that sentence ringing in their gut, that isn't a threat to your business. That's the cheapest hire you'll ever make, and you already made it.

But the direction is not in doubt, and here's what I'd actually tell you to do, which isn't what you'd expect from somebody who sells this for a living. Don't pay for a playbook. You already knew that. But don't pay for somebody else's judgment either, and that's the part most owners miss. Judgment about your shop was never going to outsource well, because you are the one who has to live with the call.

The move is to get both halves without buying either. Ask for the proposal. Let them do their work and propose the playbook, in detail, the way selling has always worked. Then run that outline against your own numbers with an AI and look hard at what comes back. The reason that output is worth taking seriously is that the AI has nothing to sell you. A consultant's recommendation is shaped, honestly and usually without anyone meaning it, by whatever that consultant happens to offer. The AI has no service line to protect. It puts the options in front of you, and then you do the part that was always yours: decide.

Worst case, you walk into the engagement understanding exactly what you're paying for. Best case, you just kept $9,500 in the business. The advantage in every services transaction has always belonged to whoever held the most context. That was never the vendor. It was always the owner. So was the judgment. The tools finally caught up.