Whito Intelligence
Proprietary data, market intelligence and advisory for the UK small business economy
Independent research, data, visibility testing and advisory for the companies that sell to UK small businesses. We record what the AI engines actually recommend, check every business against the Companies House register, and publish the method so you can attack it.
If you are a small business yourself, you do not need this. Our free research and free Scorecard are better value for you, and they always will be.
What we do
Four things. Each one exists because somebody asked us for it and nobody else in the UK was doing it.
Proprietary data
Every business an AI engine names in response to a real customer question, recorded with the engine, the town and the date, then checked against the Companies House register. The database grows every quarter and nobody else is building it.
Market intelligence
Syndicated quarterly reports, one edition per trade, covering who gets recommended, how much the engines disagree, which sources they lean on, and what moved since last quarter.
Visibility testing
We test a named brand across its locations the way a customer would, on four engines, against a competitor set you choose. You get the specific reasons for invisibility, ranked by what to fix first.
Advisory and due diligence
Custom data cuts, bespoke prompt sets, market sizing, and an independent read on a target company's real digital position before you buy it.
The rule for this section
The free research stays free. All 89 studies, no gate, no email wall, no trial. Read them at whito.co.uk/research. Whito Intelligence is not those studies with a price on. It is the named, repeated, quarter on quarter data underneath them.
Reports for purchase
One report, one price, bought outright. No subscription and no monthly fee. Each one arrives as a PDF plus the underlying data as CSV, licensed to your company. Anything not yet published carries the date it lands.
AI Recommendation Index: UK Electricians, July 2026
The first edition. Three AI engines asked to recommend a trustworthy electrician in 15 UK towns, every business named recorded, and all 182 of them checked against the Companies House register.
- 26 pages, plus the raw data as two CSV files
- All 182 businesses named, by town and by engine
- Every dissolved or liquidated company named, with company numbers and dates
- Verification split for all fifteen towns
- What the data means by buyer type, and the limits stated plainly
Sector Digital Health Benchmark
One edition per trade. A large sample of real UK firms scored against the Whito Scorecard, so you can see the true state of a market rather than the state of your own customer list.
- Share with a working website, readable prices and current reviews
- Consistency of business details across the web
- Structured data and AI readability
- Broken down by region and by company age
- Sample size stated on the front page, always
The Ghost Trader Report
The businesses AI engines confidently recommend that are dissolved, dormant or were never registered at all. The headline findings publish free, because people should know. The named data sits behind this.
- Every ghost recommendation we found, named
- Which engine produced it and on what date
- Companies House status and date of dissolution
- Breakdown by trade, region and engine
Brand and Franchise Visibility Test
The one piece of work we do to order. We test one named business across its locations, against a competitor set you choose, and report exactly where the AI engines cannot see you.
- Location by location visibility across four engines
- Named competitor benchmark, so the result has context
- The specific reasons for invisibility, ranked by what to fix first
- Board ready summary plus the full underlying data
How one edition is built
Three inputs, run on the same fixed schedule every quarter, combined into one index. Nothing is estimated and nothing is modelled.
What the data looks like
An index page ranks every business by share of AI mentions and flags the ones the register says are no longer trading.
Who buys this
Seven kinds of buyer, all of whom already spend money trying to understand this market and currently do it by guessing.
Trade software vendors
You see the market through your own customer base. This shows you the rest of it.
Lead generation platforms
Do the AI engines send people to you, or straight past you. That is now an existential question.
Franchisors and multi site brands
Find out which of your locations are invisible before your franchisees do.
Trade bodies and accreditation schemes
You tell members your badge makes them more findable. We can tell you whether that is true.
Insurers and business finance
Market sizing and segment health for a customer base that is hard to survey.
Investors and acquirers
Commercial due diligence on a target's real digital position, not the position in the deck.
Agencies serving trades
Evidence to take into a pitch, and a benchmark to hold clients against.
Something bespoke
Most of what we do is a finished report you buy outright. Occasionally somebody needs a question answered that no report covers.
Custom data and due diligence
Bespoke prompt sets, market sizing, or an independent read on a target company's digital position before you buy it. Scoped and quoted per engagement.
Talk to usSponsored research
You fund a study on a question you care about. We control the method and the findings, the funding is labelled on every page, and we publish whatever the data says. Those terms are not negotiable, and they are the reason it is worth funding.
Talk to usHow the method works, in public
If you are paying four figures for research, the first thing you should do is attack the methodology. Here is ours, so you can.
- Fixed prompts. Every prompt is written before the quarter starts and published in the report. We do not change them to make the findings more interesting.
- Multiple runs. AI engines do not give the same answer twice. Every prompt is run several times per engine and we report the variance rather than hiding it. If an engine contradicts itself, that is a finding, not a problem.
- Dated and versioned. Every result carries the engine, the date and the model version where it is available. A result from October is not compared to one from July without saying so.
- Verified against the register. Every named business is checked against Companies House. Where a name cannot be matched to a registered entity, we say so rather than guessing.
- Name matching is disclosed. Two engines rarely spell a business the same way. The matching rule is published, and so is what a stricter rule would do to the numbers.
- Sample sizes on the front page. Not in an appendix. If a sample is too small to carry a claim, we do not make the claim.
- Corrections published. If we get something wrong, the correction goes on the report and everyone who bought it gets told.
Questions people ask before they buy
What can I actually buy today?
The July 2026 electricians edition is published and delivered on purchase, and the Brand and Franchise Visibility Test is bespoke work we can start now. Everything else carries the date it lands. The quarterly programme resumes in October 2026.
Is this the same as your free research?
No. The free research is the published finding. This is the named, repeated, quarter on quarter data underneath it, with the raw responses. The free work stays free and will not be moved behind this. Read it at whito.co.uk/research.
Who can use a report once we have bought it?
One company, internally, including its own staff and its board. Quoting a figure externally is fine and encouraged, with attribution to Whito and a link. Reselling it, or passing the raw data to a third party, is not.
What if the report is not what this page says it is?
You get your money back. Digital research is normally sold with no refund at all, which we think is a poor deal. If a report does not contain what is listed on this page, email us and we will refund it in full.
Can we see a sample first?
Yes. The electricians sample is free and complete enough to judge us on. For anything else, ask and we will send the relevant sample and the full prompt set before you commit.
Do you sell the contact details of businesses in your directory?
Never, and we will not. Everything sold here is aggregate or based on publicly published information and the public register. Businesses that claim a Whito listing are not a product. If that ever changed, the directory would not be worth having.
Can a sponsor influence what the research says?
No. Sponsorship funds the work, it is labelled on every page, and we publish the findings whether the sponsor likes them or not. See how Whito makes money.
Tell us which market you need to understand
If your trade is not in the published editions, say so. The order we cover markets is decided by who asks.
Whito Ltd, registered in England and Wales, company number 10918465. All prices in GBP and exclusive of VAT, which is added at checkout where applicable.
Reports are delivered by email as PDF plus CSV, licensed to the purchasing company. Findings are based on AI engine responses recorded on stated dates and on the public Companies House register. AI engine responses vary between runs and change over time. A report describes what those engines said when we asked, and it is not a statement about the quality, competence or conduct of any business named in it.
This page describes research products. It is not legal, financial or professional advice, and you should not rely on it as a substitute for advice specific to your business. See our editorial policy, data and sources and terms.