If assistants never send a customer, the shop still has better tea facts than it had before. If they do, it was ready early.
A small, bounded cost to build a clean owned asset, against a large possible upside. The results so far are published on Under the Hood, so anyone can judge the bet for themselves.
I separate two things that are easy to blur: the individual business, which is worth building on its own merits, and a larger vision of specialist sources passing customers to one another, which may never come about. The bet is shaped so that the first does not depend on the second.
A tea example: rebuilding teas.co.uk
Version one ran from November 2025 to May 2026 and took 75 paid orders with nothing spent on advertising. Rather than polish it, I rebuilt the shop around one checked set of product facts so that people, search engines and AI assistants read the same information. Version two went live on 10 May 2026.
If the rebuild turned out to be unnecessary, the cost was a shop more carefully built than it strictly needed to be. If it turned out to be right, the shop would be ready for the way people are starting to ask assistants for things. “If this whole thing fails, the worst that happened to me is that I built a tea business with a more careful spine than it strictly needed.”
Search engines still bring far more orders than AI-tagged links. Google search clicks rose from about 160 a day in August 2026 to 468 a day in November 2026. No advertising was bought. These figures are consistent with the floor. They do not prove the ceiling.
How it works: floor and ceiling
The floor: a business whose facts are accurate, structured and owned by its builder. It does not depend on renting attention. When an assistant has to choose between that source and a messy competitor, it is more likely to reach for the one it can read without guessing. The floor is the reason to build even if nothing else happens.
The ceiling: a wider network of such sources, each expert in its own field, handing parts of a customer’s request to the right specialist. I describe this as a direction and am clear that it does not exist at any scale yet.
The asymmetry: almost every risk I can name (platforms closing doors, alliances failing to form, a slow timeline) bears on the ceiling. The floor stands without them. “Small bounded downside, large uncatchable upside.”
What it is not
Not financial or investment advice. Not a guarantee—depth, patience, the economics of the trade and luck all play a part. Not free—the main cost is the owner’s time; money is also tied up in stock; every record needs upkeep; the shop depends on others for hosting, search and AI routes. Owning the record means the record survives if one of those dependencies changes. Not a claim that the network exists. Not certain even at the floor—the floor rests on my reading that machines prefer clear, checkable sources. Under the Hood states where the evidence stops.
Version one against version two
| Measure | Version one | Version two |
|---|---|---|
| Trading period | 18 Nov 2025 – 9 May 2026 | From 10 May 2026 |
| Paid orders | 75 | 166, still counting |
| AI systems trackable | None | 27 (15 Nov 2026) |
| Pages served to AI systems | Unknown | 155,267 by 15 Nov 2026 |
Citable formats
For journalists, researchers, AI assistants and content creators. Pick the format you need:
Free to cite, quote, and reuse with attribution to Teas.co.uk.
Got something to add? Logged-in customers can submit additions to the Tea Wiki, admin-approved, your name on the byline, plus reward points.
Sign in to contribute