Guides

How ChatGPT Picks Which Melbourne Electrician to Recommend

A homeowner in Brunswick is halfway through a kitchen reno. The old switchboard won't take another circuit, the sparky who quoted has gone quiet, and it's Sunday night. They open ChatGPT and type: "Can you recommend a good electrician in Brunswick for a switchboard upgrade?"

Ten years ago that person opened the Yellow Pages. Five years ago they typed it into Google and called someone from the map. Tonight they asked a chatbot, and the chatbot gave them three names.

If you're an electrician working those suburbs, the question that matters is simple: were you one of the three? And if not, what did the other three have that you didn't?

We ran that exact search, and a few variations of it, to see how the answer gets built. Here's what's going on under the bonnet.

What actually happens when someone asks

ChatGPT doesn't hold a ranked list of electricians in its head. When you ask it for a local recommendation, it goes and looks something up, the same as you would.

It rewrites your question into a handful of its own searches first. "Electrician Brunswick switchboard" becomes several queries: "best electrician Brunswick", "switchboard upgrade electrician inner north Melbourne", "licensed electrician Brunswick reviews", and so on. This is called query fan-out, and it means your business has to match a spread of phrasings, not one exact keyword.

Those searches run against a web index. For ChatGPT that index is mostly Bing's, through OpenAI's arrangement with Microsoft, plus OpenAI's own crawler. The practical upshot: if Bing can't find you, ChatGPT usually can't either. Most tradies have checked how they look on Google. Almost none have ever looked at Bing.

Then it reads a batch of pages, picks the few it trusts most, and writes a short answer citing them. Research on how ChatGPT browses suggests most of the pages it opens never make it into the final answer. It's pulling maybe three to six sources through a fairly narrow gate.

The shape of the answer it gave

The reply we got followed a pattern that showed up every time we asked, whatever suburb we used.

It named three or four businesses. Each one got a single line of justification: this one had a large number of Google reviews at a strong rating, that one had been operating for over a decade, another explicitly advertised emergency and after-hours work. It finished with a sentence telling the person to confirm the electrician's licence and get a written quote.

Notice what those justifications are made of. Review counts. Years in business. A clearly stated service. Every one of those is a fact ChatGPT found written down somewhere public, in language plain enough for a machine to lift. Nothing in that answer came from the quality of anyone's actual wiring. It came from what the internet says about them, and how easy that was to read.

Signal 1: it can only recommend what it can find

The businesses that got named all had the same thing in common. A complete, active Google Business Profile (there's a 20-minute walkthrough of fixing yours if it's been sitting untouched). Listings on the big Australian directories. A website that said, in normal words, what they do and where.

The ones that didn't get named weren't necessarily worse electricians. Some had a Facebook page and nothing else. Some had a website that was all photos and no text a machine could use. One had a Google listing with the wrong suburb on it.

If your entire online presence is a Facebook page and word of mouth, you are close to invisible to this. There's a longer piece on where Melbourne customers actually look now if you want the full picture, but the short version: the AI can't recommend a business it can't locate.

Signal 2: your details have to match everywhere

ChatGPT cross-checks. Your business name, address and phone number on your Google profile get compared against your Yelp page, your Bing listing, True Local, Yellow Pages, your website footer, the lot.

When those line up, the model is confident it's looking at one real business. When they don't, when you're "Brunswick Electrical" on Google and "Brunswick Electrical Services P/L" on an old directory, with a mobile number that changed in 2023 still sitting on three sites, its confidence drops. A business it isn't sure about is a business it doesn't name.

This is the least exciting item on the list and one of the most common reasons a good operator gets skipped. Search your own business name and read every listing that comes back. Same name, same number, same suburb wording, everywhere. Fix the ones that are wrong, biggest sites first. The invisible-on-Google-Maps guide covers this in more detail; the same consistency problem hurts you in both places.

Signal 3: reviews, and how recent they are

Every business ChatGPT recommended had reviews it could point to. Not just a rating, a visible pile of them, and recent ones.

Third-party reviews are weighted heavily because they're the closest thing to independent proof the model has. A hundred Google reviews at 4.7, with new ones landing every month, reads as a busy, trusted business. Twelve reviews, newest one from last year, reads as either quiet or slipping, and the machine can't tell which.

Rating matters, but volume and recency do a lot of the work. An electrician steadily picking up two or three reviews a month will, over a year, look far more active to an AI than one sitting on fifteen five-star reviews that all arrived in 2024. If asking customers for reviews is the part you hate, the method that works without begging is a short text message sent the afternoon you finish the job, with the review link already in it.

Signal 4: someone other than you has to say you're good

A few of the recommended businesses turned up in "best electricians in Melbourne" style articles and curated directory pages. When ChatGPT builds a shortlist, it leans on the same best-of-city lists and directories a person would skim. Being on the reputable ones is a direct path into the answer.

You can't fake your way onto those, and you shouldn't try. But you can make sure that when a real one is being compiled, you're an easy include: claimed directory listings, a working website, reviews that back you up.

Your own site helps here too, if it answers questions in plain language. A page that actually explains what a switchboard upgrade involves, what it costs in rough terms, how long it takes, is the kind of thing that gets retrieved and quoted. A page that just says "Quality Electrical Solutions for Your Home" gives a machine nothing to work with.

Signal 5: the label a machine can read

The businesses that get surfaced consistently tend to have structured data on their website. It's a bit of code that spells out, in a format built for machines, that this is an electrical business, here's the address, here's the phone number, here's the service area, here are the hours.

You don't need to understand the code. You need to know it exists and that most tradie websites don't have it, or have it filled in wrong. It's the difference between a machine guessing what your site is about and being told. If you had a website built and nobody mentioned schema, assume it's missing.

Why this isn't the same as ranking on Google Maps

Google Maps sorts businesses into an order and shows you the top few for your location. It's a ranking.

ChatGPT isn't really ranking anything. It's deciding which businesses it's confident enough to name out loud. That's a threshold, not a leaderboard. You don't need to beat every other electrician in Brunswick. You need to clear the bar where the model trusts your information enough to put your name in front of a customer.

That's good news for a smaller operator. You're not fighting for one of three map slots. You're making sure your business is legible, consistent and backed by reviews, and plenty of established competitors haven't done that work either.

Nobody can promise you a spot in an AI's answer. Anyone who does is guessing, and it's against the rules we work under to claim otherwise. What you can do is remove every reason for the machine to skip you.

How to find out where you stand

You can test this yourself in five minutes. Open ChatGPT, or Gemini, or Claude, and ask it to recommend an electrician in the suburb you work most. Ask a few different ways. See who comes up. See if it's you.

If it isn't, the reason is almost always one of the five above, and usually more than one.

Our free AI visibility check runs that test properly, across ChatGPT, Claude and Gemini at once, and comes back with a plain read on where your business stands: whether the AI tools mention you, what your Google listing looks like to them, and which of the gaps above is costing you. Eight questions, no sales call attached.

Run the free check and see whether the chatbot in your customer's pocket knows your name.

```json

{

"@context": "https://schema.org",

"@type": "Article",

"headline": "How ChatGPT Picks Which Melbourne Electrician to Recommend",

"description": "We asked ChatGPT for a Melbourne electrician and watched how it chose. Here are the signals it actually uses, and how a tradie gets into that answer.",

"datePublished": "2026-08-27",

"author": {

"@type": "Organization",

"name": "Vexability",

"url": "https://vexability.com.au"

},

"publisher": {

"@type": "Organization",

"name": "Vexability",

"url": "https://vexability.com.au"

},

"mainEntityOfPage": {

"@type": "WebPage",

"@id": "https://vexability.com.au/blog/how-chatgpt-picks-melbourne-electrician-recommend"

}

}

```