Three names come back. Not thirty.
Your next customer is asking an AI assistant who they should be working with. It answers with two or three companies, and they go no further. We make sure your name is in that answer.
We publish our methodology. As quarterly runs complete, we publish our failure log in the open. You are welcome to check both.
“Who are the most reliable suppliers of this in my industry?”
Not thirty. Three.
65%
of Google searches now end without a click
Similarweb, 2026
94%
of business buyers used AI in their last purchase
Forrester, 2024
47%
had already built their case before contacting a supplier
Forrester, 2024
Measured across every assistant your buyers actually use
For twenty years your customers looked. Now they ask.
On Perplexity, ninety three percent of questions are answered without sending anyone anywhere.
Your buyer is not looking any more. They are asking. And the thing they are asking behaves less like a directory and more like a concierge. Ask a good concierge where to eat and you do not get a list of every restaurant in the city. You get three names, said with confidence, and you stop asking.
The question is no longer whether you can be found. It is whether you get named.
You will never see the enquiry you did not get.
When you lose a tender, somebody tells you. When you lose a pitch, someone calls. When you are left out of an AI answer, nothing happens at all. Just a quarter that is quietly worse than the last one, for reasons nobody in the building can name.
By the time your phone rings, most of the decision has already been made somewhere you cannot see.
Your website is five percent of the problem.
Most firms selling AI visibility will offer to rework your website first. It is not where the answer comes from. The rest is the accumulated record of what other people have said about you.
Think of it as the difference between your CV and your references. The CV gets you considered. Nobody in the history of hiring has been given the job on a CV alone, because the person deciding always goes and asks somebody else. AI does the same thing. It reads what you say about yourself, and then it believes what everyone else says.
So we work on the other ninety five percent. It is slower and it is harder. It is also the reason this works.
5–10%
We change what the internet says about you, in the places AI actually reads.
Find out where you stand.
We take the real questions your buyers ask, in their words rather than search terms, and we put them to every major AI assistant. Then we show you the answers. Not a summary of the answers. The answers.
Map where the answer comes from.
Every category has a small set of sources the assistants keep returning to. Directories, particular publications, certain roundups and threads. We identify yours. Most owners are surprised by how short and how unglamorous the list is.
Build your presence in those places.
Placements, listings, authoritative commentary, and contributed research where your category buyers already gather. Handcrafted editorial execution that programmatic automation cannot produce.
Keep score, in the open.
One number, measured the same way every quarter against a baseline recorded before we started. The percentage of relevant answers in which you were named. Up or down. Nowhere to hide, for either of us.
This is patient work, done by people.
An editor approached with something genuinely worth publishing. A listing claimed and corrected. A roundup from 2023 brought up to date. None of it can be automated, which is precisely why so few firms offer it and why it still works.
The method, step by stepThis is what one looks like.
A representative buying query put to a leading assistant in an industrial category. Three companies named with confidence. Any supplier missing from the answer is invisible to the purchaser.
See what AI says about youBuyer asks
Who are the most reliable suppliers of precision-machined valve components in the north of England?
ChatGPT
Based on reputation, certifications and buyer reviews, three suppliers stand out: Company Name, Company Name, and Company Name. All three hold ISO 9001 certification and are frequently recommended in industry directories and trade forums.
What it looks like when it is working.
Not traffic charts. These are the things our clients actually notice.
Enquiries arrive already convinced.
Someone who found you inside an AI answer arrives already told, by a source they trust, that you are worth speaking to. Across high-intent B2B buying cohorts, inbound enquiries from AI recommendations convert at over four times the rate of generic organic search.
You know your number.
Every quarter, one figure against the same question set, measured the same way. You will be able to say, precisely, whether you are more visible than you were ninety days ago. Almost nobody in your industry can say that today.
The answers to “how did you find us” start to change.
“ChatGPT recommended you” is a sentence business owners began hearing occasionally in 2025. The ones who did something about it now hear it weekly.
A dated baseline. A public method. A number either of us can check.
What happens, and when.
Days 1 to 10: Initial Readout
Unedited transcripts across all six assistants, raw buyer query responses, and initial category source map delivered in ten working days.
Days 11 to 30: Formal 4-Week Baseline
Multi-week repeated polling across the measurement window to eliminate prompt volatility and establish your dated Shortlist Share benchmark.
Days 31 to 90: Placement Build
The compounding work. Sources mapped, editorial citations secured, trade directories claimed and verified.
Day 90 onwards: Quarterly Score
Re-measurement against the identical baseline question set. Measured the exact same way, every 90 days.
An orchard, not a billboard. Anyone promising you a result in three weeks is selling something else.
You should be sceptical of this entire industry. So are we.
Hundreds of millions in venture funding have poured into automated AI visibility tools. Low-cost vendors charge a few hundred dollars a month to churn out AI-written programmatic articles that actively degrade domain health and harm the client they were hired to assist.
Without an objective scoreboard, every agency claims success regardless of commercial outcome.
So here is our standard. We publish our measurement method in full, including what it cannot tell you. We establish a verified, dated baseline before any placement begins. As our client programmes and quarterly research progress, we publish what worked and transparently log what did not. And on ongoing programmes, we back our work with a defined 120-day performance commitment: a minimum +15 percentage point improvement in Shortlist Share, or we work at no fee until achieved.
None of that makes us right. It makes us checkable, which in this market is considerably rarer.
Read our methodology in fullStart by finding out.
The Shortlist Audit delivers unedited transcripts and your category source map in ten working days, and your formal dated baseline at day thirty. Credited in full against your first month of an ongoing programme.
What we do not do.
- We do not build websites.
- We do not run advertising.
- We do not manage your social media.
- We do not produce content in bulk, because the volume approach is most of the reason this industry has the reputation it has.
We do one thing, and we would rather be the best in the world at it than adequate at nine.