The AI Opportunity Map · For founders of software and AI companies
Their customers have already written your AI brief.
We read 25,773 low-rated Google reviews of UK local businesses to find which failures AI could take over, where they happen most, and what to build first.
Reviews to 30 October 2026 · One report · One-off payment · Instant access
★☆☆☆☆Not reachable“couldn't get through to them after trying for several days”
Garages & MOT
★☆☆☆☆Updates and follow-up“chasing them for updates because they never communicated anything”
Estate agents
★★☆☆☆Booking failures“rescheduled or cancelled at the last minute”
Dentists
★☆☆☆☆Charges and refunds“nearly twice as much as my original quote”
Removals
The headline finding
Almost a third of bad reviews describe a failure AI could take over.
A failure is a customer who could not reach the business, heard nothing, was let down on a booking, or was surprised by a bill. Only 13% describe one with no complaint about staff or the work as well, so the report keeps the two apart.
Used car dealers fail customers more than twice as often as the average trade
Failures per 1,000 reviews at independent firms, at most 50 reviews per firm. Trades run from 6 to 49. The report ranks all 18, with likely ranges, share of bad reviews and UK firm counts.
In some trades, most of the trouble is process
That share of bad reviews at estate agents describes these failures. At cafes and takeaways it is 2%.
The starting point differs
That share of independent garages has no website, against 5% of vets. A website tool misses most garages, so phone and text come first.
Few show an AI tool yet
Of the 3,980 homepages we read, that share names an AI chat or voice tool in its code. AI phone answering does not show on a homepage, so treat this as a minimum.
In the customer's own words
The brief is already in the reviews.
Local business owners buy relief from a specific, daily problem. Their customers describe it in public, often in more detail than a discovery call.
“They never answer the phone when I call”One star review · Estate agents
“said one price and then charged me more”One star review · Garages & MOT
“always promising to get back to you and never do”Two star review · Used car dealers
“They cancelled me on the same day”One star review · Removals
Word for word from public Google reviews of independent firms, with businesses and people left out. The report sets out four quotes for each group of failure.
What AI will not fix
A report that tells you where not to pitch.
Quality of the work and staff attitude come first in bad reviews, and AI fixes neither. A pitch that promises more will not survive a read of the firm's own reviews.
What one and two star reviews are about
Share of 7,657 reviews of independent firms. A review can mention several things, so these overlap. The report covers every theme.
Inside the report
Seven sections, from the finding to your first line.
- Summary
The headline figures, and what they do not mean.
- What customers complain about
Every theme in one and two star reviews, showing what AI can and cannot fix.
- Trade ranking
All 18 trades, with likely ranges, share of bad reviews and UK firm counts.
- The failures
Four groups of failure, each with quotes, what to build, what to measure and how far to trust the label.
- What firms already have
Gaps on Google by trade, and the tools found on 3,980 homepages.
- Where to start
Three trades worked through, each with a first build and an opening line to adapt.
- Method and limits
Where the figures come from, how the AI was tested, and what the report cannot tell you.
- Get the report
One of the three opening lines
Lead with their problem, not your portfolio.
Each trade in the Where to start section comes with a line like this, built on what its customers write most often.
An opening line to adapt
“Reviews of local garages keep coming back to bills higher than expected and to customers ringing to ask if the car is ready. We set up a written price before work starts and a text at each stage.”
Who it is for
- Founders of software and AI companies choosing which local trade to sell to
- Agencies that want a specific first line in place of a generic pitch
- Anyone scoping an AI product for local businesses
What it is not
- A list of leads: it names no businesses
- A forecast of budgets: it measures customer need, not intent to buy
- A figure for all UK businesses: it covers these 18 trades
Full report
The AI Opportunity Map
£15
One-off payment · Instant access
Questions
Before you buy.
- Where does the data come from?AI read 25,773 low-rated reviews, out of 292,800 public Google reviews in English, comparing 6,489 independent firms. We also use Office for National Statistics business counts and our own read of 3,980 homepages.
- How accurate is the AI reading?We tested it on 300 reviews that a second AI model read separately: 89% precision and 86% recall across the five failures. Some labels are weaker, and the report says which. No person has audited the labels.
- Which trades does it cover?Used car dealers, estate agents, hotels, garages, removals, tyre shops, vets, wedding venues, accountants, dentists, restaurants and pubs, roofers, physios, aesthetics clinics, electricians, barbers, cafes and driving schools.
- What does it leave out?Buying signals, regional rankings, trends and market sizes in pounds. Solicitors, pharmacies, opticians, care homes and cleaners have no reviews here.