# The complete payload: everything a tea shop hands an AI assistant in one answer

**Canonical URL:** https://teas.co.uk/ai/complete-payload/
**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

The complete payload is Lee Samuel Tucker's name for the full, structured answer a business hands an AI assistant in a single trip: what the product is, what it costs, whether it is in stock, the proof behind each claim, its limits, and the route to buy it.

## Description

The complete payload is my name for the full, structured answer a business hands an AI assistant in a single trip: what the product is, what it costs, whether it is in stock, the proof behind each claim, its limits, and the route to buy it. I group it into four parts: product truth, claim boundaries, a citation graph and an action layer.
When an assistant asks about a tea, the shop that answers in one go, with the price, the stock, the proof and the route to buy, is the easiest one to recommend.
Ordinary pages make an assistant take many separate trips, and every trip is a chance for the facts to disagree. The complete payload turns those dozen trips into one answer: what the tea is, what it costs, whether it is in stock, the proof, its limits and the way to buy it.
The complete payload is the practical side of a truth entity. I use the term to describe the difference between an assistant having to piece together an answer from a page made for human eyes and an assistant being handed a prepared, checkable answer by the business itself. At teas.co.uk, the independent tea shop I run from Tunbridge Wells, it is the reason products carry structured data behind their ordinary pages.
First, what is a payload?
A payload is the useful information carried in a software response. For a tea-shop question, it might include the matching products, their pack sizes, prices, caffeine status and links. It is the contents of the reply, rather than the envelope used to deliver it.
My phrase “the complete payload” describes the aim of returning the information needed for a decision together. It does not mean every product record contains every possible fact, or that every question can always be answered in one request.
Four parts in ordinary language

Term
What it means for a shopper

Product truth
The recorded facts about the exact item: which tea, which pack, price and availability.

Claim boundaries
What can honestly be said about it, including limitations and unknowns.

Citation graph
The links between statements and the evidence supporting them.

Action layer
The permitted next steps, such as opening the product page or building a basket when asked.

An API can return this information as JSON, text organised into named values and lists. A longer explanation can also be supplied as Markdown, a plain-text format with headings and links. The formats should preserve the relevant facts and limits from their authoritative source.
A tea example: fewer trips, one agreed record
Imagine someone asks an assistant for a strong loose-leaf breakfast tea, no bags, under ten pounds. Working from ordinary web pages, the assistant has to make many separate trips: one for the page, one for the price, one for the picture, another to find a checkout link, with every trip a chance for one fact to disagree with another. It may end up with a low-confidence guess and no source.
As I put it: “The complete payload turns those dozen trips into one.” Where the question allows, the assistant can ask once and get the product, its price and stock, the facts it needs to judge the request, the sources behind those facts and a link to buy, all drawn from the same record the shop’s own page uses.
The more a customer says, the more this matters. A full request of the kind I call Premium Intent, with its context, constraints and budget, can only be met if every fact it depends on comes back together and agrees.
The scenario is illustrative. The structure behind it is what the shop is built around.
The four parts

Product truth. Exact names, values with their units, the price and the stock position.
Claim boundaries. What a product is, what it is based on, and what it should not be bought for.
Citation graph. Each claim linked to the supplier document or other source behind it, so a reader can check it.
Action layer. The checkout link and the rules that say where everything resolves, ready for an assistant to act on with the customer’s permission.

My longer description adds context around those four: the brand, relevant Tea Wiki and recipe pages, the main product image and its records, related products, and a short way for the assistant to report whether the route worked.
What is built today
Built and live: every product on teas.co.uk has structured data behind its page, including identifiers, prices, availability, source links and price per cup, all read from one checked set of product facts introduced with version two on 10 May 2026. Source links and other details appear where the record holds them, and missing ones are marked as gaps. Since 6 October 2026, an assistant can, with the customer’s approval, search, compare and fill a basket through the teas.co.uk MCP server, the shop’s connection for assistants built on the open Model Context Protocol. The customer checks the basket and pays on teas.co.uk, and the assistant never sees their card details.
The extra fields in my longer description are the full design. Not every one is returned for every product on the live site yet, so the list above describes the aim rather than a promise about any single product.
A useful answer can still have gaps
A dated check of the Tetley Decaf Pure Green record on 15 November 2026 returned its identity, decaffeinated status and related resources, while marking the response as usable with known gaps. Some attributes were unclaimed or needed a pack check. The honest answer includes those limits instead of pretending “complete” means flawless.
Fetching several pages is not inherently wrong, and a normal product page can already include several facts. Bringing relevant facts together can reduce extra requests; the exact number depends on the question and the available service. Checkout and stock should still be checked at the point of purchase.
What the complete payload is not
It is not a scrape of a web page, and it is not marketing copy in a different format. It does not hand over payment: the customer pays on the shop’s own checkout. It does not oblige an assistant to use it or to cite the shop.
I am open about a risk here. The companies running assistants could take a clean answer and serve it inside their own walls, without ever sending the customer or the credit back. My answer is that live stock, the checkout and the shop’s accountability for its own claims stay with the shop, and that clean outside sources are cheaper for an assistant than building everything itself, so it still has reasons to route the customer there. That is an argument, not a result. The shop’s Under the Hood pages publish what it can and cannot see of that journey.
Where the term comes from
The idea came from my work rebuilding teas.co.uk in 2026. I started calling it the complete payload in June 2026 to describe what the rebuilt shop should hand an assistant when it asks about a product, and to name the gap between an answer pieced together from pages and an answer the business prepares itself. I don’t know whether anyone has used the name before, but it suits what I have built.
Quick reference

Detail
Information

Term
The complete payload

Meaning
A full, structured, sourced answer about a product, brought together so an assistant needs as few trips as possible

Used by
Lee Samuel Tucker, founder of teas.co.uk

Used at teas.co.uk since
June 2026

Opposite or contrast
An assistant scraping a page made for human eyes and guessing

Related
Truth entity, claim boundaries, Premium Intent, built versus planned

What others have found
Outside research and reporting on the wider subject. It is background: it does not test or prove my own terms.
Google’s guide to product structured data and the schema.org Product type describe the machine-readable facts, such as price and availability.

---

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