Do AI Assistants Recommend Your Melbourne Business? How to Check — and How to Fix It

Ask ChatGPT to recommend a web designer, a physio or an accountant in Melbourne and it will name actual businesses — confidently, in a sentence or two, with no page-two to scroll to. If your business isn't in that answer, you don't rank lower. You don't exist.

The numbers behind that are blunt. In 2026 testing across thousands of local queries, ChatGPT recommended only 1.2% of local businesses and Perplexity about 7.4% — against roughly 36% visibility for the same businesses in Google's local 3-pack. Meanwhile Google's AI Overviews now appear on an estimated 30–40% of searches, and ChatGPT fields around 200 million weekly users asking exactly the kind of conversational, "who should I use" questions that used to start on Google.

That gap — businesses easy to find on Google but invisible to AI assistants — is the single biggest free traffic opportunity for Melbourne small businesses right now. Here's how to measure where you stand, and what actually fixes it.

The 10-minute self-test

Open a private browser window (so your history doesn't flatter the results) and put the same five questions to ChatGPT, Perplexity and Google:

1. "Who's the best [your category] in [your suburb]?"
2. "I need a [your category] in Melbourne — who do you recommend and why?"
3. "Is [your business name] any good?"
4. "[Your business name] reviews"
5. "Compare [your business name] with [your main competitor]"

Score each answer three ways: were you named, was the information about you accurate, and who was recommended instead of you. Most Melbourne businesses fail question 1 and 2 outright, and — more worrying — get outdated or plain wrong details on question 3. That's your baseline, and it's also the exact list of gaps to close.

Why AI assistants skip you (it's not your Google ranking)

AI assistants don't crawl-and-rank the way Google's classic results do. They synthesise an answer from sources they trust, and each platform trusts differently:

• ChatGPT leans heavily on encyclopaedic and directory-style sources — one 2026 analysis put Wikipedia at nearly half of its citations — plus licensed data and Bing's index.
• Perplexity loves community and editorial sources: Reddit alone accounts for almost half of its local citations, and it strongly favours recent content — pages under 30 days old earn about 3.2× more citations.
• Google AI Overviews builds on Google's own index, enriched with authority and structured-data signals.

Three practical consequences. First, being talked about matters as much as being linked to — unlinked brand mentions in directories, reviews and community threads feed AI answers. Second, your data hygiene is load-bearing: business profile accuracy on ChatGPT and Perplexity sits around 68%, so assistants routinely quote wrong hours, old addresses and dead phone numbers unless the public record is spotless. Third, reviews are a gate: businesses recommended by ChatGPT average about 4.3 stars, and testing suggests businesses under ~3.5 stars with unanswered reviews are effectively invisible.

The seven fixes, in order of leverage

1. Make your NAP identical everywhere. Name, address, phone — character-for-character the same on your site, Google Business Profile, Bing Places, Yelp, industry directories and socials. Inconsistency reads as unreliability to a model deciding whether to cite you.
2. Get your reviews above 4.3 and answer them. Volume matters less than average and responsiveness. Reply to every review — response rate is a visibility signal, not just good manners.
3. Add structured data (schema) to your site. LocalBusiness, Organization, Service and FAQ schema tell machines exactly what you do, where, and for whom. It's a 20-minute job for a developer and most Melbourne sites still don't have it.
4. Publish direct answers, with dates. AI assistants quote pages that answer questions in the first two sentences, cite specific numbers, and show a recent date. A pricing page that says "In 2026, a business website in Melbourne typically costs A$3k–10k" is citation bait; "contact us for a quote" is invisible.
5. Add an llms.txt file. It's a plain-text index at yoursite.com/llms.txt that tells AI crawlers what your site contains and which pages matter. Adoption is only around 10% of sites, so it's still an easy differentiator.
6. Get mentioned in the places assistants read. Curated directories, local press, industry round-ups and genuinely useful Reddit answers. One honest, detailed community answer about your specialty can outweigh a dozen backlinks nobody reads.
7. Keep pages fresh. Perplexity's recency bias means a quarterly refresh of your key pages (with an updated date stamp) quietly compounds your odds.

The Melbourne angle

Local intent is where AI recommendations are most decisive — "near me" style questions produce short, confident shortlists, not ten blue links. That cuts both ways: there are fewer winners, but the winners take almost everything. For most Melbourne categories the AI shortlist is still soft — built from thin directory data and a handful of reviews — which means a few weeks of disciplined fixes can put you in answers your bigger competitors assume they own.

That's also why we publish this directory the way we do — structured data on every profile, dated updates, llms.txt, plain-language answers. It's the same playbook we're recommending to you, running in production.

Want the audit done for you?

We run a free AI Search Visibility Audit for Melbourne businesses — we test how you show up across ChatGPT, Perplexity and Google AI, check the seven factors above against your web presence, and send you a prioritised fix list. It's valued at $497, it's genuinely free, and there's no sales call attached. Request it via the popup on this site, or get matched with a web team if your site needs the deeper rebuild first.

FAQ