By Susan Jones, Co-Founder, WebMax Canada

Picture handing a customer a quote with a confident, clean-looking number on it, only to find out later the math behind it was wrong. Not wrong in a way that was obvious. Wrong in a way that sounded completely right until you re-checked it by hand and the total didn’t hold up. Now you’re either eating the difference or having an awkward call explaining why the number changed.

That’s the specific way AI tools have already burned a few contractors and trades owners we’ve talked to, and it’s worth naming plainly before getting to the part where these tools are actually useful, because both things are true and the order matters.

One owner we met on the road this spring, somewhere between BC and Alberta, put it in a way that stuck with us. He’d tried letting an AI tool help with the math side of his estimates once. It got the figures wrong, confidently, and he caught it before it went out. After that, he made a permanent rule for himself: the AI gets the words, it never gets the numbers. He didn’t read that rule anywhere. He learned it the way trades learn most things, by testing the tool on a real job and finding out exactly what it was good for, the hard way, once.

Most contractors and trades owners are already using AI tools more than they let on. Not for anything dramatic, just emails, cleaning up the wording on a quote, a quick explainer on something outside their specialty. That’s not a trend prediction. That’s just where things already are.

The harder question isn’t whether to use these tools. It’s knowing where they’re reliable and where they’re not, and that line is clearer than it sounds once you’ve tested it on real work, ideally before it costs you a customer’s trust.

Writing and rewording is the strongest use case by a wide margin. If you’ve got a rough, technically accurate quote that reads like it was typed on a job site at 6pm, because it was, an AI tool can turn that into something a customer reads as professional without you spending twenty minutes on phrasing. The substance stays yours. The polish gets faster.

The same goes for emails, simple explainers on materials or processes you don’t deal with often, and a first draft of anything you’d otherwise stare at a blank page over. Think of it as a fast first draft you still read before it goes anywhere, not a finished, trusted document.

Math. Specifically, math that ends up on an invoice or a material order. Square footage calculations, quantity takeoffs, cost totals. These tools can sound completely confident while being wrong, and a wrong number that looks right is more dangerous than an obviously broken one, because nothing about it makes you stop and double check.

The fix isn’t complicated: keep AI tools on the writing side of any document and do your own math, or run any AI-generated number through a second check before it goes to a customer. That’s not a workaround for an immature technology. It’s likely to remain the right rule even as these tools improve, the same way a calculator never replaced the habit of estimating a number in your head first to sanity-check the output.

Picture it as two different tools in the same kit. One is excellent at translation, turning a rough idea into clean language fast. The other one you wouldn’t trust yet, because you’ve already seen it be confidently wrong. Most owners who’ve spent real time with these tools land on roughly that same split without anyone explaining it to them, because it shows up clearly the first time a calculation goes sideways.

That instinct, test it on a real job, learn the boundary the hard way once, then respect it permanently, is a more reliable approach than either blanket enthusiasm or blanket distrust. Both extremes miss what’s actually true: the tool is genuinely useful and genuinely fallible, and a sensible business owner plans around both facts at once rather than picking one to believe.

If you’re talking to anyone about using AI to help your business get found by customers, the same logic applies one level up. Be skeptical of guarantees. A vendor who promises a guaranteed outcome from an AI tool is making the same mistake as trusting AI with your invoice math: treating a probabilistic, constantly changing system as if it were a fixed, reliable one. The honest version of this work is built on solid information and consistent details that can be checked, not on guarantees that any one tool will behave a certain way next month.

The same gut check that saved your invoices is worth applying to whoever you hire for this. If a marketing company hands you a confident-sounding report about your visibility and you can’t tell what it’s actually based on, that’s the same wrong number that looked right. Ask what it’s checked against. A vendor who’s done this work honestly can show you the real listing, the real review, the real page, not just a number that sounds good. You learned not to trust a confident answer you couldn’t verify. That instinct is worth keeping when you’re the one being quoted, not just when you’re doing the quoting.

Susan Jones is Co-Founder of WebMax Canada, a 100% Canadian-owned digital marketing company based in Victoria, BC, and Google Certified, helping contractors and trades businesses across the country build a clear, consistent online presence for over 10 years. See more on how AI is changing how customers find local businesses on the WebMax Canada blog.

Explore related articles you won’t want to miss:

You Can’t Get Prequalified the Day the Tender Closes

If You Can’t Read the Quote in One Sentence, Ask Again

• 
•