Last Updated on July 10, 2026
Can you trust ChatGPT, Gemini or Google AI to recommend an electrician? This is Whito’s original UK research, July 2026 data. We asked all three tools for a trustworthy electrician in 15 UK towns. They named 182 different businesses, only 6% were named by all three, and just under half could be verified as active registered companies. In 6 of the 15 towns, at least one tool recommended a business whose registered company was dissolved or in liquidation.
More and more people are not searching Google for a tradesperson. They are asking ChatGPT, Gemini and Google’s AI. “Recommend a trustworthy electrician near me.”
So we did what those customers do. We asked all three tools the exact same question in 15 UK towns, logged every business they named, and then checked every single one against the Companies House register.
What came back should worry every electrician in the country. And it explains a problem most of them do not know they have.
The mistake: assuming AI knows your business exists
Ask an electrician how they get found and they will talk about Google, reviews, maybe a leads site. Almost none of them have thought about the tools that are quietly becoming the first place customers ask.
That is the mistake. Being invisible to AI is the new version of not being on Google. It happens silently, and by the time you notice the drop in enquiries, you have been losing them for months.
What we found
We asked ChatGPT, Gemini and Google’s AI Mode the same plain question, “Recommend a trustworthy electrician in [town]”, across 15 towns from Sunderland to Plymouth. That is 45 answers to what is essentially one question.
Those 45 answers named 182 different businesses.
Only 11 of the 182, about 6 percent, were named by all three tools. Three quarters were named by one tool and one tool only. In 8 of the 15 towns, the three tools could not agree on a single electrician between them.
That lands almost exactly on our earlier finding across trades, where AI engines agreed on the best tradesperson just 8 percent of the time. Same test, different trade, same answer. The tools are not drawing on some shared, reliable picture of who is good. Each one is guessing from whatever it can read.
Then we checked the recommendations against the official register, and it got worse.
Just under half of the businesses named, 49 percent, could be matched to an active registered company in or near the town they were recommended for. The other 51 percent could not. Most of those simply do not appear on the register at all under the name the AI gave. Some will be legitimate sole traders, who do not have to register. But a customer cannot tell which is which, and neither can the AI.
And in 6 of the 15 towns, at least one tool confidently recommended a business whose registered company had already been dissolved or was in liquidation. One was described as an established local electrician with a 4.9 rating. Its company was struck off more than two years before we asked. Another was dissolved just weeks before our test and still came back as a trustworthy recommendation.
One tool even recommended a firm for Luton whose registered company sits in Staffordshire, and the tool’s own map card pointed at an appliance shop 26 miles away in another county. Same name, wrong business, presented with complete confidence.
The tools could not even agree with themselves. During the test, Gemini showed us two different candidate answers to the same Nottingham question side by side. The two lists shared one name out of nine.
Why this happens
AI tools do not think. They pattern match against the information available to them. To recommend an electrician with any confidence, a tool needs clear, consistent, verifiable signals that the business exists, does that work, and serves that area.
Most electricians give it almost nothing to work with. Their details differ from site to site. Their business is not verified against any official record. Their trading name does not match their registered name. So the tool does what it always does when it is unsure. It guesses. Sometimes the guess is a competitor. Sometimes it is a company that no longer exists.
The electricians who do show up are not the biggest or the best. They are the ones whose information is clearest and easiest to trust.
What this means for your business
If customers are starting their search by asking AI, and the AI cannot confidently find you, you are losing work before you ever hear about it. There is no missed call to notice. The enquiry simply goes to whoever the tool named instead. In our test, half the time that was a business no one could check, and occasionally it was a business that no longer exists.
This is happening now, not in some future. The share of customers who ask an AI tool first is only going one way.
The good news is that this is fixable, and most of your competitors are not paying attention yet. Being early here is a real advantage.
How to get found by AI
Make your business easy for a machine to trust. That is the whole task.
Be verified against an official record, so a tool can confirm you are a real, registered business rather than a name on a page. If you trade under a different name from your registered company, say so in plain text where both a person and a machine can read it.
Keep your details identical everywhere. Same business name, phone number and service area, with no variations. Inconsistency reads as uncertainty, and uncertainty gets you left out.
Be specific about what you do and where. “Electrician” serving “the UK” is too vague to recommend. “NICEIC registered electrician serving Leeds and LS postcodes, domestic and commercial” gives a tool something solid to work with.
Gather reviews and keep them current. They are a trust signal a machine can read as easily as a human.
None of this is clever or expensive. It is structure. And structure is exactly what the machines reward.
The methodology, so you can check us
In July 2026 we asked ChatGPT, Gemini and Google AI Mode the same question, “Recommend a trustworthy electrician in [town]”, once each in 15 UK towns: Reading, Sunderland, Swansea, Leeds, Nottingham, Coventry, Luton, Bristol, Doncaster, Ipswich, Preston, Stoke-on-Trent, Milton Keynes, Norwich and Plymouth. We used the free tier of each tool with default settings, because that is what a real customer sees. ChatGPT switched to its lower-quality fallback model partway through, which is also what a real customer sees.
We logged every business named, 237 mentions covering 182 unique businesses, and searched each one on the Companies House public register. We counted a business as verified when an active registered company matched the name in or near the recommended town. Not appearing on the register does not make a trader fake, because sole traders do not have to register. That is exactly the point. A recommendation you cannot check is a recommendation you are taking on faith, and so is the AI.
Where a tool offered two alternative answers, we recorded the first. This is a snapshot of one question type in one month, not a census, and results will vary by tool, phrasing and time. We have not named the dissolved or mismatched firms, because the failure here belongs to the tools doing the recommending, not to small businesses.
The sharp takeaway
Customers are already asking AI which electrician to trust. Across 45 answers the tools named 182 different businesses, agreed on 6 percent of them, could not verify half of them, and in 6 of 15 towns recommended a company that had ceased to exist.
The AI is guessing. And right now it is probably not guessing you.
The electricians who win the next few years will not be the ones with the flashiest vans. They will be the ones a machine can find, verify and confidently recommend. That is not luck. It is structure, and you can put it in place before your competitors even realise the game has changed.
Frequently asked questions
Can you trust ChatGPT to recommend an electrician?
Treat it as a starting point, not a verdict. In this test, just over half of the recommended names could not be verified as active registered companies, though some will be legitimate sole traders. Check any AI recommendation against Companies House and the Registered Competent Person Electrical database before you hire.
Will AI recommend your electrician business?
Only if a machine can find and trust you. The businesses AI named tended to have consistent details across the web, current reviews and a verifiable official record. If your details vary from site to site, or your business cannot be matched to an official record, you are unlikely to be the name AI gives a customer.
Whito verifies UK electricians against Companies House and structures each listing so people and AI tools can read and check the detail. Claim your free listing and start being findable by the tools your customers now ask first.
