How AI Search Is Changing How Australians Find Tradies (And What Tradies Need to Do)

Five years ago, a homeowner needing a plumber in Brisbane typed "plumber near me" into Google. Today, more of them are asking ChatGPT "who's the best emergency plumber in Coorparoo and how much should it cost?" The answer they get isn't a list of ten blue links. It's a paragraph summarising 2 or 3 businesses, with their reasoning, their hours, and sometimes their phone number, all delivered before the homeowner ever visits a website.

This is generative engine optimisation, or GEO. It's the next phase of search marketing for local trades, and it's already changing how Australian tradies get found. The rules are different to traditional SEO, the optimisation moves are different, and most tradies haven't started.

What's actually changing in homeowner search behaviour

Three shifts are happening in parallel. None of them is hypothetical. They're all visible in current Australian search data.

First, Google's AI Overviews now appear above the traditional search results for a growing share of "find me a tradie" queries. When a homeowner searches "concreter Bulimba", they often see a Google-generated summary recommending businesses before they see the map pack. The summary draws from websites Google's AI considers authoritative for that query.

Second, ChatGPT (with web search), Perplexity, Claude, and Gemini are increasingly used as discovery tools, not just chat tools. Homeowners ask "find me a roof restoration company in Wynnum" the same way they'd ask a friend, then act on the answer. The conversational format means there's no scrolling through ten options. Usually 2 or 3 businesses get named.

Third, Apple, Samsung, and Google are baking AI assistants into phones at the OS level. "Hey, find me an EV charger installer in Newstead" is becoming a voice query answered by an AI agent. The agent picks one business to surface first, sometimes only one.

The combined effect: where 10 tradies used to compete for visibility on page one of Google, 2 or 3 now compete for the AI's recommendation. The winner takes more. Everyone else gets less.

How AI engines pick which tradies to recommend

AI search isn't magic. It's a structured retrieval process. When ChatGPT or Google AI Overviews answer a tradie query, they're combining three signals.

Signal one is content the AI can read and understand cleanly. That means well-written copy on your website, with clear structure (headings, paragraphs, bullet points), specific information (what trades you do, where you operate, what you charge in ranges, what licences you hold), and machine-readable markup (schema.org structured data telling the AI exactly what the page is about).

Signal two is authority signals beyond your own site. That includes Google Business Profile reviews, mentions on other Australian websites (local news, directories, suppliers), and how often other sites in your trade link to or reference you. The AI is essentially asking: does the rest of the internet treat this business as legitimate?

Signal three is intent matching. Does your site clearly answer the question being asked? A site that says "we paint houses" won't beat one that says "we paint weatherboard houses in Ashgrove, Bardon, and Paddington, fully licensed and insured, typically $4,800 to $9,500 for a full exterior". Specificity wins.

This is why hyper-local trade sites tend to do well in AI search. A site like painter Ashgrove or concreting Morningside is built to match exactly one intent for exactly one area. When ChatGPT is asked about painters in Ashgrove, a single-focus site beats a generalist tradie site every time.

The new rules for being findable by AI

The good news is that the rules favour smaller, focused businesses over big aggregators. AI engines actively try to avoid recommending shared lead platforms because the answers aren't useful. "Try HiPages" isn't a recommendation, it's a deflection. AI engines learn to skip those answers and go to the underlying business directly.

That said, your site needs to make the AI's job easy. Most tradie sites don't. Three common failures.

Failure one: thin content. A homepage with "Quality plumbing for Brisbane families since 2018" tells the AI almost nothing. It can't extract a phone number, a service area, a specialisation, or a price point. Compare that to a page with clear headings ("Areas we serve in Brisbane North", "Standard call-out fee", "Licensing and insurance details"). The second is citable. The first isn't.

Failure two: no schema markup. Schema.org structured data is the layer of machine-readable information that tells AI engines exactly what your business is, where you operate, what you charge, what licences you hold, and what reviews you have. Most tradie sites have none of it. The handful that do gets cited first.

Failure three: no llms.txt. A new file specification (llms.txt) has emerged in the last year, sitting in your site root alongside robots.txt. It tells AI crawlers in plain English what your business is, what it does, and how it should be referenced. Tradies who add this file get clearer, more accurate AI citations. Tradies who don't get omitted or misrepresented. See leadlord.com.au/llms.txt for an example.

What you can do today: a 7-point AI-search checklist

Why this matters more for trades than for most industries

Most AI search optimisation advice is written for software companies and ecommerce. Trades are different in two ways that make AI search uniquely powerful.

First, tradie queries are local and urgent. A homeowner with a burst pipe doesn't want to read seven blog posts. They want one trustworthy answer in 10 seconds. AI search delivers exactly that, and rewards businesses set up to provide the structured answer the AI needs.

Second, trade purchasing is high-intent. There's no browsing. If the AI recommends you, the homeowner usually rings within minutes. That makes AI search higher-converting than most traffic sources. A small share of AI-recommended traffic is worth more than a large share of generic SEO traffic.

The flip side is the failure mode is much harder. If you're not the recommendation, you don't appear at all. There's no page two of an AI answer. Either you're cited, or the homeowner never hears your name.

What this means for tradies in 2026

Two things are happening in parallel and reinforcing each other.

The first is that traditional SEO is becoming less reliable. Google has compressed its search results page to make room for AI Overviews. The map pack that used to dominate is now competing for space with a generated paragraph. Click-through rates from organic results to tradie sites have dropped measurably across most trades in the last 18 months.

The second is that AI search is more concentrated. Fewer businesses get recommended per query. The winner takes a much larger share of the available enquiry volume than the average ranked-3 result on traditional Google.

The result is a barbell market. Tradies who optimise for AI search well get more leads at higher quality. Tradies who don't get progressively fewer leads at lower quality. The middle is shrinking.

The good news, if you're starting now, is that most tradies haven't worked out what to do yet. The infrastructure (schema, llms.txt, structured content, intent-specific pages) is straightforward and cheap. The competitive window for being early is open right now and won't be for much longer.