AI picks products. It reads your product data, not your website.
We find what actually decides a purchase in your category, make sure your products say it, and keep improving it against what sells.
Book DemoHere are three that hold up on a commute.
Compared on
- needsnoise handling type · not stated
- needsear tip sizes · not stated
What changed
Your website is optimised. Your Amazon listing is optimised. ChatGPT is not.
Your website
Written for
The campaign that sent the click
Persuade someone who is already standing in front of the product.
Your Amazon listing
Written for
The Amazon ranking
Win the slot against the listing sitting next to it.
ChatGPT/Gemini
Reads
Both of the above
Compares you against everything else, on whatever the buyer just asked.
You are sending thin data made for human persuasion to an AI deciding on hundreds of criteria.
And you cannot see it happening
No error. No report. No drop to trace.
A product that cannot be matched to a question simply stops being shown. Nothing anywhere reports it.
No error
Nothing flags a product that cannot be matched to a question.
No report
No dashboard shows a recommendation you were not part of.
No drop to trace
Traffic does not fall. It never arrives in the first place.
Your product stops being shown and nothing tells you. That is why it has to be checked every month.
What we make
Your catalogue is Transaction Ready. Is it Buyer Ready?
For years a catalogue had one job: help someone buy, by giving them the minimum needed to understand the product and place the order. A Buyer Ready catalogue also carries what someone needs to judge whether it is right for their situation.
Today
Transaction Ready
Answers: what is this?
- What the product is, and what it costs
- Enough to understand it and place the order
- Written for someone who already chose to look
What we make
Buyer Ready
Answers: who is this for?
- Which situations it suits, and which it does not
- The details a buyer self-selects on, stated plainly
- Readable by a machine putting an answer together
What decides a purchase
Most listings lead with what matters least
The details that decide a purchase are weighted, and the weighting is different for every product type. We hold that order. Then we check your products against it, one product type at a time.
How it works
Then we find out which answers actually sold
We already hold what buyers settle on in your category. The first month checks your catalogue against it, answers what is missing and puts it live. After that we watch which of those details actually produced sales, and do more of what did.
One record. Each one gets it in the format it reads.
Impressions
CTR
AI sessions
AI revenue
Same brand, same period, same demand. The difference is which products answered the question.
Why this is still broken
Nobody writes and measures in the same place
Your feed tool, your tracking tool and your agency each do one part of this. None of them does the part next to it, so nothing ever learns from what sold.
Feed tools
Publish reliably. They move whatever you give them, and have no view on what a buyer needed the product to say.
Visibility trackers
Report what happened. They can tell you that you were left out, and have no way to change the product data that would have got you in.
Agencies
Write well. They produce content for people, and cannot see which detail got the product into a ChatGPT answer.
Writing without measurement is guesswork.
Measuring without being able to change anything is just a report. Neither one gets better on its own. Doing both, on the same catalogue, is the only thing we do.
Where it goes
One record. Every assistant reads it.
One Buyer Ready record per product, published into the feeds these assistants read when they put an answer together.
Your own site, your listings and your marketplaces read the same record. See where else it gets used.
See where your catalogue stands
We will run a readiness check on one of your products and show you exactly what is missing, at no cost.
Book Demo