Home / Proof
New name. Not new at this.
The proof, before the case studies.
There are no client case studies under this name yet, because the company is new. Rather than pad the gap with stock logos and invented testimonials, this page shows the actual evidence we do have and explains exactly how you can check it.
The number, and what it is not
Over a million citations. Not impressions.
An impression means a page appeared somewhere. A citation means the answer engine picked your page as the authority worth naming. The second one is the entire product, and it is the only one we report.
The number
AI and Bing citations across the sites we look after. Not impressions, not views. Citations, where the machine names a source as worth quoting.
Screenshots, dated, on the call. We are not going to ask you to trust a number printed on our own website.
Why we are showing our own sites, not client logos
Because the alternative would be made up.
Every new agency faces the same moment: no case studies, and a wall of blank space where the client logos are supposed to go. The usual solutions are a stock photo of a handshake, some invented testimonials, or a strip of logos belonging to companies you once did a small job for.
Faking the proof would be a strange way to open that relationship, so here is the real position instead.
What we have: a decade of this work, a portfolio of our own properties that AI systems demonstrably cite, and a mechanism we can explain step by step because we run it on ourselves every month.
What we do not have yet: a named client under this brand with a published before and after. That is exactly what the free first month is for, and it is why the free seat requires signed permission to publish results, whatever they turn out to be.
When those case studies exist they will replace this page rather than being added to a pile. One real thing at a time.
The mechanism, in public
This is the loop we run every month.
Not complicated. Just nobody does it for businesses your size. Click a step.
Run the prompt set across four engines
Fifteen prompts, four engines, five runs each, logged out, fixed location, same day every month. Five runs because a single answer is noise, not data.
Record which domains it actually cited
Not just whether you appeared. Which specific pages the machine leaned on to build its answer. This is the step everyone skips.
Go and get you onto those exact sources
Directories, review platforms, industry roundups, the specific publications the model already trusts. Plus the content on your own site that gives it something worth quoting.
Then run the whole thing again
Same prompts, same engines, same day. The score moves or it does not, and the cited-source list changes, which tells us where to point next month.
We publish this because knowing the method is not the hard part. Doing it every month for twelve businesses without dropping any of them is the hard part.
What we will not claim
The honest limitations.
- We cannot guarantee you will show up in ChatGPT. Nobody can. Anyone who does is guessing, and you should ask to see that guarantee in writing.
- It is slower than the website. The site is live in seven days. Traffic and enquiries usually move within a couple of months. Being named in AI answers is slower and depends heavily on your category.
- If your category already has a dominant player, we may not be able to move it, and we will tell you that on the first call rather than six months in.
- The website alone is not the lever. Around nine in ten AI answers cite third-party pages. If the off-site work does not happen, a rebuild on its own will not move the number, which is precisely why the off-site programme is the spine of the retainer and not a bullet point.
Run it on your business.
On the call we run your real numbers across all four engines, with screenshots, whether or not we end up working together.
See if you qualify