How AI search engines decide which business to recommend

When someone asks ChatGPT, Perplexity, or Google's AI search, 'Who should I hire for kitchen renovation in Vancouver,' the AI doesn't guess. It reads.

It scans your website. It checks your Google Business profile. It reads your reviews. It looks at how you describe your work. And it decides: does this business actually do kitchen renovations in Vancouver, or just claim to.

If you're vague, it skips you. If you're specific, it recommends you.

This is called answer engine optimization—making sure AI search engines understand what you do and find you when customers ask. Unlike Google's old search, which counted links and keywords, AI search reads for proof.

What AI search actually looks for

When an AI search engine decides which business to recommend, it checks four things.

First: does your website say you do this work. Not in marketing speak. In plain English. A dental clinic in Kelowna that says 'We offer comprehensive oral health solutions' is invisible. A clinic that says 'We do root canals, cleanings, and emergency extractions' gets found. The AI is looking for the actual words a customer would use.

Second: is there proof. Reviews are proof. Case studies are proof. A portfolio with before-and-after photos is proof. Marketing copy is not. When an AI reads that you've done 200 kitchen renovations, it checks your reviews to see if customers say the same thing. If they do, you rank higher. If they don't, the AI notes the gap.

Third: how recent is the proof. A five-star review from 2019 is weaker than a three-star review from last month. AI search engines assume that recent work is current work. If your website hasn't been updated in a year, or your last review is from two years ago, the AI assumes you might not be active.

Fourth: how specific are you. A plumber in Richmond who lists 'plumbing services' ranks lower than one who lists 'drain cleaning, water heater replacement, and leak repair.' A property manager in Surrey who says 'we manage properties' ranks lower than one who says 'we manage 40 to 100-unit residential buildings in Surrey and Delta.' Specificity tells AI that you know what you're doing and that you've done it before.

The gap between what you think you're saying and what AI reads

Here's what happens in practice.

You have a website. It looks good to you. But when an AI search engine reads it, it sees:

• A homepage with a photo and a tagline, but no list of what you actually do
• A services page that uses industry jargon instead of customer language
• A portfolio with beautiful images but no description of the problem you solved
• Contact information that's hard to find or incomplete
• No address, or an address that doesn't match your Google Business profile

Meanwhile, your competitor down the street has a simple website that says exactly what they do, where they do it, and what customers say about them. When an AI search engine compares the two, your competitor wins.

You're not losing because your work is worse. You're losing because the AI can't see your work clearly enough to recommend you.

Three places where most businesses lose the recommendation

Missing or incomplete Google Business profile. This is where AI search engines look first. If your phone number is wrong, your address is missing, your hours are outdated, or you haven't posted in six months, the AI assumes you're not actively taking work. A 12-person software company in Gastown might think their Google profile is just for local search. It's not. It's one of the first places AI checks when deciding whether to recommend you.

Reviews that don't match your claims. An accounting firm in Burnaby says 'we specialize in tax planning for tech founders.' But the reviews mention bookkeeping, payroll, and GST filing. The AI notes the mismatch. It will still recommend you, but lower, because the proof doesn't match the claim. If you want to rank for 'tax planning,' you need reviews that mention tax planning.

Website language that customers don't use. You know what you do. Your team knows what you do. But when a customer searches, they use different words. A distributor in Calgary might call their service 'supply chain optimization.' Customers call it 'getting parts on time.' If your website only uses your language, AI search won't connect you to the customer's question.

Where to start

• Audit your Google Business profile this week. Check that your phone number, address, hours, and website URL are correct. Add a recent photo. Write a two-sentence description of what you do in plain English. This is where AI search engines look first, and it takes 15 minutes.

• List your actual services in customer language. Open your website's services page. Rewrite it so a customer (not your team) would understand it. Use the words customers use when they search. A plumber should list 'drain cleaning' not 'hydro-jetting.' A clinic should list 'root canals' not 'endodontic treatment.' This takes 30 minutes and moves the needle.

• Ask your last three customers to leave a review, and mention what they hired you for. When a review says 'they fixed my leaking tap quickly,' that's good. When it says 'they diagnosed and replaced my mixing valve in 45 minutes,' that's better—because it's specific. Specific reviews help AI search understand what you do. Send a text or email asking for a review, and suggest they mention the specific problem you solved. This takes 10 minutes and takes a week to see results.