When AI Doesn’t Know Your Business Exists, Neither Do Your Customers

ByTech Support
Home service businesses across Alberta are quietly losing jobs to AI recommendation tools that only pick one contractor per search. Here’s what’s actually driving the drop — and why your reviews and rankings won’t show you the problem.

Somewhere in the last year, a contractor in Calgary stopped getting the calls he used to get. Not all of them. Just the good ones — the kind of job that used to fill a slow Tuesday.

His reviews were fine. Four and a half stars, dozens of them, people saying nice things about the crew. His rankings hadn’t moved. His website traffic was flat, not falling. Nothing in any dashboard he owned explained where the work went.

Here’s what he couldn’t see: someone typed a question into Google Maps — need someone who can look at a leak this week, decent reviews, not overpriced — and Google’s AI picked a name. His wasn’t it. There was no runner-up list, no page two to eventually get found on. Just one recommendation, maybe two, and everyone else got nothing. Not a lower ranking. Nothing.

That’s the part that should actually worry you. Losing a bid is normal business. Losing a bid you never knew existed is a different problem, and there’s no line item for it.

Why Google’s AI Only Recommends Up To Three Local Business

Google’s doing this on purpose. The feature is called Ask Maps, a Gemini-powered tool that launched across the US in March and replaced the old scroll-through-ten-results habit with a straight answer to a specific question.

New research from SOCi put a number on how selective that AI actually is: across nearly 350,000 business locations, ChatGPT recommended only 1.2% of them.

Gemini did better at 11%, Perplexity landed around 7.4%. Compare that to Google’s traditional local three-pack, where the same businesses showed up 35.9% of the time. Restaurants had it worse — 83% of them didn’t show up in AI recommendations at all.

This is layered on top of a search environment that was already sending fewer people anywhere. SparkToro’s Rand Fishkin found that 68.01% of Google searches in early 2026 ended without a single click to any website — up from 60.45% just two years earlier.

When an AI Overview shows up on a search, people click through only 8% of the time, according to Pew Research, compared to 15% when it doesn’t appear. The web traffic funnel that used to feed local businesses their leads has been narrowing for years.

The AI recommendation layer is what’s replacing the wide end of it, and it’s built to pick one winner, not present ten options.

For a high-ticket trade — roofing, HVAC, foundation work, renovation — that’s not a marginal loss. That’s the $18,000 job that never became a phone call.

Calgary, Edmonton, Red Deer: The Same Problem, Three Businesses

None of this comes down to who does better work. It comes down to who left the AI something to go on.

A plumber in Edmonton has the certifications, the trucks, the fifteen-year track record. His Google profile hasn’t been touched since it was set up. No new photos in two years. Reviews from customers never get a response. To a person, that business is clearly running. To an AI scanning for signs of life, a quiet profile reads as a business that might not be — and it moves on to the next listing without flagging the mistake to anyone.

A renovation company outside Red Deer has the opposite issue. The profile exists, the reviews are strong, but half the fields are empty. No listed hours. No services breakdown. No description beyond the business name. When someone asks an AI for a renovator who does kitchens and is available this month, the AI has nothing to match against the question. AI isn’t ranking that business lower. It’s leaving it out of the conversation entirely.

And back in Calgary, the contractor with the great reviews has a subtler problem: his customers keep writing “great guys, fair price” instead of the specific words people actually search for. If nobody’s review ever mentions “same-day,” the AI has no way to connect his fifteen years of same-day work to a query asking for exactly that. The skill is real. The evidence, in the AI’s eyes, isn’t.

Three different businesses, three different gaps, the same result — invisible to a system that only recommends what it can verify.

Why the Skilled Ones Are Getting Skipped

The businesses that do get picked tend to share a pattern: recent activity, specific language in their reviews, and a profile with no blanks in it. In the SOCi data, Culver’s hit a 30% recommendation rate on ChatGPT and 45.8% on Gemini, largely on the strength of consistent ratings and a fully built profile. Businesses with ratings near 3.4 stars and almost no review responses were, in the report’s own words, effectively invisible.

Estimates vary, but multiple studies put the share of local businesses with incomplete, unclaimed, or mismatched Google listings somewhere between 40% and 55%. That’s not a small technical oversight. It’s a large share of Alberta’s home services sector handing the AI a half-finished form and expecting to still get picked.

This Isn’t an SEO Problem

This isn’t a general SEO problem, and it isn’t solved by posting more, in the abstract. It’s narrower and more specific than that — and most owners have never opened the one profile this actually depends on since the day someone set it up for them.

The gap between the businesses getting recommended and the ones getting skipped isn’t talent. It’s whether anyone ever went back and finished the job on the digital side of the business the way they’d never leave a job half-finished on-site.

Why This Keeps Getting Worse

The zero-click number isn’t leveling off. Google’s AI Mode passed a billion monthly users by I/O this year, and query volume is more than doubling every quarter. Whatever share of the market AI recommendations control right now, it’s the smallest it will ever be from here.

The contractor in Calgary eventually noticed the calls had changed. The plumber in Edmonton hasn’t yet. The renovator in Red Deer probably won’t, until someone tells him what an AI sees when it looks at his listing — which is, at the moment, almost nothing at all.

Fixing this isn’t complicated, but it does take someone that knows what AI is looking for when it scans a business profile. We work with Alberta home services companies to close exactly those gaps.

Sources:

  • SOCi 2026 Local Visibility Index — industry press coverage
  • SparkToro, “In 2026, Less Than One Third of Google Searches Still Send a Click” — https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
  • Search Engine Land, “Google zero-click searches reach 68% in early 2026” — https://searchengineland.com/google-zero-click-searches-2026-study-479717
  • Pew Research Center, AI Overview click-through study, 2025
  • Google I/O 2026 announcement, AI Mode usage figures
  • BrightLocal Local Consumer Review Survey, 2026