There's a test I've started running in almost every client meeting.
I open ChatGPT, type something like "what's the best [type of business] in [city]" and hit enter. Then I turn the screen toward the business owner sitting across from me.
Most of them have never done this before. They assume the results will be like Google — familiar names, their own probably somewhere in the mix. What they actually see lands differently. A competitor they've been outperforming for years gets named as "the go-to choice." A business that closed eight months ago gets a glowing recommendation. Their own company — one they've spent a decade building — isn't mentioned at all.
The room gets very quiet.
I've run this experiment across dozens of local businesses over the past year. Dental clinics, law firms, healthcare providers, real estate agents, medical spas. The results are different every time, but the pattern is the same: AI has already formed an opinion about your business — and you probably have no idea what it is.
Here's what I found when I tested ten of them.
The Setup
For this exercise, I ran each business through the same three prompts across ChatGPT, Google Gemini, and Perplexity:
- "Who is the best [business type] in [city]?"
- "I'm looking for a [business type] near [neighbourhood]. Who do you recommend?"
- "What do people say about [Business Name]?"
I'm keeping the businesses anonymous — the point isn't to embarrass anyone, it's to show you the patterns. I've changed identifying details but kept the errors real.
What I Found at a Glance
| # | Business Type | City | AI Engine | Error Type | Business Impact |
|---|---|---|---|---|---|
| 1 | Dental Clinic | Toronto | ChatGPT | Stale specialization | Attracting wrong patient type |
| 2 | Personal Injury Law Firm | Vancouver | Gemini | Orphaned contact info | Dead-end phone number |
| 3 | Medical Spa | Calgary | Perplexity | Underweighted reviews | Lost consultations |
| 4 | Real Estate Team | Toronto | ChatGPT | Stale specialization | Wrong leads, missing right ones |
| 5 | Physiotherapy Clinic | Ottawa | Perplexity | Citation platform gap | Ranked below inferior competitors |
| 6 | Accounting Firm | Montreal | All three | Citation invisibility | Zero AI presence after 11 years |
| 7 | Hair Salon | Toronto | ChatGPT | Corrupted data point | Wrong service expectations |
| 8 | Dental Specialist | Mississauga | Gemini | Missing brand narrative | Competitor given false advantage |
| 9 | Fitness Studio | Hamilton | Gemini + Perplexity | Citation volume gap | Chains recommended instead |
| 10 | Family Law Firm | Toronto | Perplexity | Ghost employee | Credibility undermined at intake |
What I Found
Business #1: A Dental Clinic (Toronto)
The error: ChatGPT described this clinic as specializing in "general family dentistry." They pivoted to a cosmetic and implant-focused practice over two years ago. Invisalign is now their primary service. Their Google Business Profile, website, and every review platform reflects this — but AI still describes them the old way.
Why it happened: The clinic's early online presence was built around family dentistry. That language lived in old directory listings, a Yelp profile from 2019, and a cached version of their previous website. AI learned from that data. The newer positioning simply hasn't propagated into the sources AI trusts.
The cost: Every time a high-value cosmetic patient asks AI for an implant specialist in their area, this clinic doesn't come up. The patients it does attract via AI are often price-sensitive family dentistry seekers — the opposite of their target client.
Business #2: A Personal Injury Law Firm (Vancouver)
The error: Gemini recommended this firm when asked about personal injury lawyers — but cited a phone number that was disconnected. The firm had moved offices and changed their main line fourteen months earlier.
Why it happened: The old number was embedded in dozens of legal directories, bar association listings, and citation aggregators. AI scraped and learned from those sources. The firm updated their website immediately after the move but never did a systematic citation audit across the 40+ platforms where their old details lived.
The cost: Every referral AI sent their way hit a dead end. No call back. No explanation. Just a disconnected number — which, to the caller, reads as unprofessional or out of business.
Business #3: A Medical Spa (Calgary)
The error: When asked "what do people say about [Business Name]," Perplexity surfaced a negative review from 2021 and summarized it as representative of the client experience. The business had 94 five-star reviews since then. The AI ignored them entirely.
Why it happened: The negative review lived on a high-authority platform (RealSelf) that AI citation engines weight heavily. The volume of positive reviews was concentrated on Google — a platform AI often underweights for sentiment synthesis compared to specialized review sites in the healthcare space.
The cost: A prospective patient in the research phase asks AI about the clinic, gets back a summary that implies subpar care, and books elsewhere. The clinic never knows the consultation was lost.
Business #4: A Real Estate Team (Toronto)
The error: ChatGPT recommended this agent as a specialist in "downtown condo sales." She moved to the suburbs three years ago and now focuses exclusively on family homes in Mississauga and Oakville.
Why it happened: Her early career was built downtown and well documented — press mentions, sold listings on aggregators, a profile in a real estate publication. AI recognized her as a downtown condo expert because that's where her digital footprint was deepest and most cited. Her suburban pivot simply hasn't built the same citation density yet.
The cost: Inbound leads from AI are looking for something she no longer offers. Meanwhile, the buyers and sellers she actually wants to work with don't see her mentioned at all in their market.
Business #5: A Physiotherapy Clinic (Ottawa)
The error: This one was subtle. Perplexity mentioned the clinic accurately but described it as "a solid mid-tier option" compared to two competitors it called "top-rated." The clinic actually outperforms both on every measurable metric: Google rating, review volume, wait times.
Why it happened: The two "top-rated" competitors had more structured citations in healthcare-specific directories — Healthgrades, RateMDs, Zocdoc — that AI systems disproportionately weight in the health sector. This clinic relied almost entirely on Google reviews. More reviews, better rating, lower AI authority.
The cost: Patients who use AI to compare options before booking see a ranking that doesn't match reality. The clinic loses conversions to competitors it's objectively better than.
Business #6: A Boutique Accounting Firm (Montreal)
The error: Not mentioned at all. Zero. Across all three AI engines, across all three prompts, in a city where they've operated for eleven years.
Why it happened: The firm had essentially no structured digital footprint outside their own website. No directory listings beyond a basic Google Business Profile. No press coverage. No mentions in business publications. No LinkedIn activity. No client testimonials on third-party platforms. To AI, they were invisible — not because they're bad, but because they'd never given AI anything to learn from.
The cost: An entire channel of potential clients — people who discover service providers by asking AI — doesn't know this firm exists. This is the most common situation I encounter.
Business #7: A Hair Salon (Toronto)
The error: ChatGPT recommended the salon for "natural hair and braiding services." The salon doesn't offer braiding. They specialize in color, cuts, and extensions. The confusion appears to stem from a Google Maps edit submitted by an unverified user.
Why it happened: AI citation engines pull from aggregated data sources, including user-generated content and crowd-sourced edits. A single bad data point, if it lives on a high-authority platform, can corrupt how AI describes a business.
The cost: Clients arrive expecting a service that isn't offered. The salon wastes chair time on consultations that go nowhere. Worse, the AI-recommended clients who actually needed braiding leave disappointed — and leave reviews reflecting that disappointment.
Business #8: A Private Dental Specialist (Mississauga)
The error: When asked to compare this specialist to a named competitor, Gemini described the competitor as having "more modern facilities." There is no factual basis for this. The specialist opened in a newly built clinic two years ago; the competitor's space is over fifteen years old.
Why it happened: The competitor had been featured in a local business journal profile that described their "state-of-the-art" space — language written years ago, still indexed and cited. The specialist's new facility had never been covered in any publication or described in detail anywhere AI could access.
The cost: Patients comparing options get AI-generated misinformation that actively favours a competitor — not because of anything real, but because a years-old puff piece exists and their own story doesn't.
Business #9: A Fitness Studio (Hamilton)
The error: The studio was accurately described — but only by ChatGPT. Gemini and Perplexity didn't mention it at all, defaulting to larger chain options despite the studio having far more local reviews and a stronger local reputation.
Why it happened: Chain gyms have massive national citation footprints. They appear in every directory, every city guide, every listicle. A local independent studio — even one that genuinely outperforms them — can't match that citation volume without a deliberate strategy.
The cost: Users who ask AI for local gym recommendations and want a boutique experience get pointed to chains. The studio, despite being exactly what those people are looking for, doesn't exist in the AI's answer.
Business #10: A Family Law Firm (Toronto)
The error: When asked about the firm by name, Perplexity returned a summary that included the name of a lawyer who left the firm two years ago — positioned as a current senior partner.
Why it happened: The former partner was well-documented across legal directories, bar association listings, and media mentions. After they left, the firm updated their own website but didn't systematically clean up third-party mentions. The departed lawyer also didn't update all of their own listings, some of which still referenced the old firm.
The cost: Prospective clients research the firm, see a name they may have been referred to — and then can't find that person when they call. It creates confusion, undermines credibility, and sometimes costs the firm the intake entirely.
So Why Does This Keep Happening?
Every error above traces back to the same root cause: AI engines don't experience your business the way your customers do.
They don't walk through your door. They don't read your latest Google reviews. They don't notice that you updated your website last month or that your phone number changed in Q3.
What they do is pull from a vast, distributed network of data sources — directories, review platforms, news sites, social profiles, forum mentions, citation aggregators — and synthesize that into a picture of your business. That picture was formed over time, weighted toward sources with the most authority, and it updates slowly.
Your Google Business Profile matters. But so does RateMDs, Healthgrades, Avvo, Yelp, Houzz, Justia, Zocdoc, Trustpilot, Crunchbase, local business journals, LinkedIn, Reddit threads where someone mentioned you three years ago, and a hundred other places you've probably never thought about in the context of AI.
If those sources are incomplete, outdated, or contradictory — AI builds a picture of your business that's incomplete, outdated, or wrong.
And here's the part that should concern you: that picture is being served to high-intent buyers right now, without your knowledge, and you have no mechanism to correct it unless you build one.
The Five Failure Patterns
Based on running this test across dozens of businesses, these are the failure modes I see most often:
| # | Pattern | What It Looks Like | Most Common In |
|---|---|---|---|
| 1 | Stale specialization | AI describes what you used to do, not what you do now | Post-pivot or rebrand businesses |
| 2 | Orphaned contact info | Old phone, address, or email on directories you forgot about | Businesses that have moved or restructured |
| 3 | Underweighted review platforms | Strong Google reviews, weak presence on sector-specific directories AI trusts | Healthcare, legal, home services |
| 4 | Ghost employees | Former team members still associated with your business | Any business with staff turnover |
| 5 | Citation invisibility | Not enough structured mentions across authoritative sources for AI to cite you | Most small and mid-size businesses |
The most fixable failure mode is also the most overlooked: citation invisibility. You don't need to outrank anyone — you just need to exist in the sources AI trusts.
The Five-Minute Test You Should Run Right Now
Open ChatGPT, Gemini, and Perplexity. Run these three prompts — substituting your own details:
"Who is the best [business type] in [your city]?"
"I'm looking for a [business type] in [your neighbourhood]. Who do you recommend?"
"What do people say about [your business name]?"
Read the results carefully. Ask yourself:
- Is your business mentioned?
- Is it described accurately?
- Is the information current?
- Are competitors recommended instead of you?
What you find is your current AI reputation — the one being served to potential clients before they ever visit your website.
What To Do If You Find Problems
The good news: every failure mode above is fixable. The bad news: fixing them requires knowing exactly where your gaps are across dozens of platforms — which takes time to audit manually.
The fastest way to get a complete picture is our free AI Visibility Audit. It shows you:
- How your business appears across ChatGPT, Gemini, and Perplexity
- How you compare to your top competitors
- Where the specific gaps in your citation footprint are
- What to fix first
It takes 60 seconds to run and tells you more about your AI presence than most businesses have ever known.
Free · No credit card · Results in 24 hours
See How AI Describes Your Business
Find out if ChatGPT, Gemini, and Perplexity are sending you clients — or sending them somewhere else.
Run Your Free AI Audit →Lorne Fade is the Founder & CEO of Fade Digital, Canada's leading Generative Engine Optimization agency. He has spent 20+ years at the intersection of search algorithms and brand visibility, and now helps businesses build citation authority in the AI engines that are replacing traditional search. Connect on LinkedIn or read more from Lorne.