# Meta Building: How One Person Built a Tea Shop With AI

**Canonical URL:** https://teas.co.uk/ai/meta-building/
**Source:** teas.co.uk, UK tea specialist, Tunbridge Wells, Kent

## Shop teas.co.uk from an AI assistant (MCP)

teas.co.uk sells tea, herbal and fruit infusions, hot chocolate and Border biscuits.
MCP server: https://teas.co.uk/mcp (Streamable HTTP), listed in the official MCP Registry as teas.co.uk
Backup address (same server, same tools; use it if /mcp is blocked): https://teas.co.uk/wp-json/teas/v1/shop-mcp
Server card: https://teas.co.uk/.well-known/mcp/server-card.json
Tools: find_products, get_product, compare_products, find_recipes, delivery_and_returns, quote_cart, get_promotions, report_problem, find_substitutes, start_subscription, join_teas, ask_lee, tea_page, search_tea_wiki, read_wiki_article, brand_page, teas_home, browse_shop, browse_recipes, recipe_page, teas_news, read_news, under_the_hood, teas_socials, teas.
Basket: add_to_basket, view_basket, update_basket, checkout_link, checkout_session, basket.
With sign in: get_return_status, get_invoice, get_profile, my_orders, track_order, reorder, my_subscriptions, change_subscription, rewards_balance, start_return, my_account.
No sign in is needed to search, compare, build a basket or get a checkout link. The person always pays on teas.co.uk.
Cards: in assistants that draw MCP Apps cards, products, comparisons, the basket, checkout, orders, repeat deliveries (pause, resume, cancel), reward points and returns appear as cards.
Every assistant except ChatGPT, until OpenAI approves the next version: quote_cart, get_promotions, get_return_status, get_invoice, report_problem, find_substitutes, start_subscription, join_teas, ask_lee, tea_page, search_tea_wiki, read_wiki_article, brand_page, teas_home, browse_shop, browse_recipes, recipe_page, teas_news, read_news, under_the_hood, teas_socials, teas, basket, my_account.
Set up guides: https://teas.co.uk/ai/

## Summary

Building something with AI is fast. The risk is that it is also fast at going wrong quietly. Meta-building is the deliberate step back that stops that…

## Description

Building something with AI is fast. The risk is that it is also fast at going wrong quietly. Meta-building is the deliberate step back that stops that happening: after each piece of work, asking whether it did what was meant, not just whether it ran.
Meta-building is my name for the habit of checking the work while doing it: keeping the aim in view, inspecting what actually changed, correcting direction and writing down the decisions so the next session starts from the truth. It is how teas.co.uk has been built with AI coding agents.
One person built this tea shop with AI tools, and checking every step is what kept the facts about each tea right.
Building something with AI is fast. The risk is that it is also fast at going wrong quietly. Meta-building is the deliberate step back that stops that happening.

Metacognition, in plain words
Psychologists call noticing and checking your own thinking metacognition. It is an established word, and I do not claim it. Meta-building is the same habit applied to a build: noticing what you assumed, checking it against what is really there, and changing course when the two do not match.

The loop

Build one piece.
Inspect what actually came out, on the live site, not what the tool says it did.
Explain any gap: why is this value missing, why does this page disagree with that one?
Correct the source of the problem, not just the symptom.
Record the new state, the evidence and what is still unknown in a handover note.

In the book’s version of the loop, the step after inspecting is a question: does this piece still serve what the shop is for? For teas.co.uk that purpose is plain. A person, or an AI assistant acting for them, should be able to find the right tea, see true facts about it and buy it. A change that runs but makes a tea harder to find, or puts a wrong pack size in front of a shopper, has drifted from that purpose, even when nothing looks broken.
The handover note holds the loop together. An AI model starts every session knowing nothing about the day before, so the written note is its memory, and mine as well. Each note says what the project is for, what exists, what has been tested and what happens next. I later recognised the idea from the one-minute meeting on the Gap shop floor, where everybody starts the shift with the same picture in their head.
A shop example
The instruction is “add decaf to the product filters”. A quick build adds the filter and moves on. Meta-building keeps going. Which products does the filter now show? Are any decaffeinated teas missing, and if so, why: is the caffeine status blank in their records, or recorded in a different way? Does the filter keep decaffeinated and caffeine free apart? Does the live page show the same answer as the product records? Then the handover note says what was changed, what was checked, which records still need a caffeine status and what to do next.
The example is illustrative. The point is the order of work: data first, then the rendered result, then the written record.
Why it matters for a shop
A shop that asks people and AI assistants to trust its facts has to know when its own facts have drifted. Meta-building is the working habit behind the shop’s single source of truth and behind the rule of built versus planned: you cannot say honestly what is built unless you have checked it.
More people ask an AI assistant to find, compare and buy, instead of scrolling through pages of results. An assistant can only recommend a tea well if the facts it reads are right: the bag count, the price, whether the tea is decaffeinated, whether it is in stock. An independent tea shop that keeps those facts checked can be the source an assistant relies on. A shop that lets them drift hands the assistant a wrong answer to pass on.
The shop has never paid for advertising. Its case, to shoppers and to assistants, rests on facts that hold up when somebody checks them, and meta-building is how those facts are kept that way.
What it is not

Not a claim that AI does the thinking. The person whose name is on the shop stays responsible for the result.
Not a guarantee. Checks miss things, which is why a failed check is written down, not hidden.
Not a new science. It is a name for a careful working habit that many good builders already have.

Seeing the end first
Meta-building starts with the finished result, not the first step. I picture the end in detail, then work backwards: what has to be true just before that, and what has to be true before that.
For a tea shop, the finished result is a shopper getting a right answer. Picture someone asking an assistant for “an everyday black tea in a big box that costs under 5p a cup”. Working backwards from a good answer gives a chain of things that must already be true:

The assistant names a real pack that is in stock, such as Typhoo Premium Everyday, 240 tea bags, at .
Just before that, the assistant must be able to read a price per cup, not only a pack price.
Before that, the shop must publish the price per cup in a form software can read, kept in step with the one product record.
Before that, the product record must hold the correct bag count and the current price.
Before that, somebody must have checked the bag count against the pack itself.

Each line in that chain becomes a piece of work, and each is checked before the next begins.

---

_Content available under teas.co.uk citation contract. AI training: yes. Search: yes. Answer-input: yes._
