One record · three rails · kept in sync

When AI does the shopping, your data does the selling.

Show up for every question your products can answer. We fill in what your product data is missing and publish it to all three agent rails — with a receipt on every order it earns you.

Founding pilot: free for selected merchants
~13×
AI-referred orders YoY (Shopify, Q1 2026)
AI-referred visits & orders, Q2 2026 vs Q2 2025 (Shopify)
2.5×
AI sessions landing on a product page vs traditional search
3
Rails, one record: feed · JSON-LD · agents
PLATFORMShopify Native
SHOPIFY_APP_STOREApproved & published
CLOUD_PROGRAMGoogle for Startups
BUILT_ONGoogle Cloud · Vertex AI
MERCHANT_CENTER_OAUTHGoogle-verified
PRODUCT DATA FOR AI COMMERCE

Get your products into the AI answers

Bids decide the auction. Attributes decide entry.

Shoppers ask the agents now, and agents read structured data. UCP Fluent turns the prose on your product pages into the fields every AI surface expects, then writes the receipts back into your orders.

Apply for the founding pilot See a sample record
SCROLL
HOW IT ALL WORKS

Your product’s journey to every AI surface

Keep scrolling. Every step below is the same bottle.

01 WHAT YOUR PRODUCT HAS TODAY
Illustration — generic product
YOUR BRAND Insulated Steel Bottle
$34 In stock
THE RECORD AS IT STANDS 8 FIELD TYPES
DESCRIPTION

Keeps drinks cold all day. Double-wall vacuum insulation, 18/8 stainless steel, dishwasher-safe lid, 500 ml.

4 FACTS TRAPPED IN PROSE — NOT MACHINE-READABLE
STRUCTURED FIELDS
Barcode (GTIN)
Material
Weight
Google category
Q&A
02 WHAT UCP FLUENT ENRICHES IT WITH

Same product, same page — but every field an agent looks for is now filled in. When a shopper asks for something your product actually is, it can finally come up.

ENRICHED
YOUR BRAND Insulated Steel Bottle
$34 In stock
THE SAME PRODUCT, AS A MACHINE-READABLE RECORD 5 BLANKS FILLED
DESCRIPTION

Keeps drinks cold all day. Double-wall vacuum insulation, 18/8 stainless steel, dishwasher-safe lid, 500 ml.

THE SAME 4 FACTS — NOW LIFTED OUT INTO FIELDS
STRUCTURED FIELDS ⓘ
Barcode (GTIN)0505623901234FILLED
Material18/8 stainless steelFILLED
Weight320 gFILLED
Google categoryHome > Kitchen > Drinkware > Water BottlesFILLED
Q&A30 pairsFILLED
WHAT THESE FIELDS ARE
Barcode (GTIN)The global trade number. It is how Google, marketplaces and agents recognise this exact product across every surface — and what GTIN-gated placements require.
MaterialWhat the body is made of. Feeds filters, comparisons and "is it plastic?" answers.
WeightShipping weight, and the number an agent quotes when a shopper asks how heavy it is.
Google categoryGoogle’s own taxonomy node (GPC brick). A missing or wrong category quietly limits where you can appear at all.
Q&AQuestion/answer pairs an AI surface can quote as-is instead of inferring an answer from your prose.
And more — the fields your record never had HOVER OR CLICK TO OPEN
03 AND THE FIELD TYPES THAT WERE NEVER THERE
8 → 23 FIELD TYPES

Same product, same page — hover a section to see what each field is and what it does.

ATTRIBUTES ⓘ
capacity500 ml
dimensions7.5 × 7.5 × 26 cm
insulationdouble-wall vacuum · 24 h cold
caredishwasher-safe lid, hand-wash body
product_typeDrinkware > Water Bottles > Insulated
conditionnew
custom_label_0bestseller
WHAT THIS SECTION IS
capacityVolume in the unit each surface expects. This is what "500 ml bottle" queries match against.
dimensionsHeight × width × depth — what "will it fit a cup holder?" is answered from.
insulationThe performance claim, structured: type plus hours cold and hot, instead of a sentence.
careWashing instructions split by part, so an answer can be exact about lid versus body.
product_typeYour own taxonomy path. Powers site search, Shopping filters and agent narrowing.
conditionnew / refurbished / used. A required feed field — missing it can disqualify the listing.
custom_label_0Your private bucket for campaign or merchandising logic, e.g. bestseller.
MERCHANDISING ⓘ
product_highlight×6
“Cold for 24 hours” · “Fits standard cup holders”
product_detail×12
Materials : Body : 18/8 stainless steel
Care : Lid : Dishwasher-safe
written back to Shopifyagentic description · metafields · JSON-LD
WHAT THIS SECTION IS
product_highlightUp to six short benefit lines Google can show verbatim under your listing.
product_detailsection : attribute : value triples — your spec table in a form machines can filter on.
written back to ShopifyThe agentic description and catalog metafields we save onto the product, plus JSON-LD on the storefront page. Everything stays in your store.
GOOGLE’S 6 CONVERSATIONAL ATTRIBUTES ⓘ BUILT FOR AI MODE
question_and_answer30 PAIRS
Will it keep water cold a full workday?
— Yes, 24 hours, per the listing.
Can it go in the dishwasher?
— The lid can; hand-wash the body.
item_group_titleInsulated Steel Bottle — 3 colors
variant_optionColor: Sage · Charcoal · Sand
popularity_rankshipped
related_productshipped
document_linkshipped
WHAT THIS SECTION IS
question_and_answer30 pairs written from your own copy. This is what AI Mode reads out when a shopper asks a question in words.
item_group_titleThe family name that groups your variants into one product instead of three unrelated listings.
variant_optionThe option axis and its values — Color: Sage · Charcoal · Sand — so a surface can offer the right one.
popularity_rankA relative demand signal inside your own catalog; helps a surface pick which of your products to show.
related_productLinks to the products that belong beside this one, for bundles and "goes well with" answers.
document_linkManuals, spec sheets and care guides an agent can cite when a question goes deeper than the page.
8 → 23 field types · +100 searchable terms · uncertain values never ship
CLOSE
04 OUT OF YOUR STORE, DOWN THREE RAILS
Your Shopify store
one enriched catalog · 23 field types
RAIL 1 ⓘ Google’s feed Merchant Center API
WHAT MAKES THIS RAIL DIFFERENT

What it buys you: every Google shopping surface at once — Shopping ads, free listings, AI Overviews and AI Mode all read this one record, and the same enrichment lifts the paid and the unpaid together. It is the only rail that carries the six conversational attributes AI Mode reads.

A structured feed pushed to Google through the Merchant Center API. Google defines every field, so a missing or malformed one quietly limits which queries you are eligible for — in Google’s own words, providing all relevant attributes “increases its chances of showing for the most relevant search queries.” Google also grades what you send: a disapproved listing drops off every surface above — the next section shows how we watch and fix those verdicts.

RAIL 2 ⓘ The crawler rail on-page JSON-LD
WHAT MAKES THIS RAIL DIFFERENT

What it buys you: the assistants Google’s feed never reaches — ChatGPT, Perplexity, Claude, Gemini — reading your products straight off your own page. Nothing to submit, no approval queue, and it is live the moment the page updates.

The same facts are published as JSON-LD on your product page, so anything that crawls the web can read them there. The trade: it can only ever show what is genuinely on the page.

RAIL 3 ⓘ Shopify’s agent rail catalog · MCP · ACP
WHAT MAKES THIS RAIL DIFFERENT

What it buys you: the order without the visit. A shopper can find your product and buy it inside the conversation, so you are sellable to people who never open your site.

Your catalog published into Shopify’s own agent plumbing — the Shopify Catalog and your store’s MCP endpoint — which is what feeds the Shop app, Shopify’s global catalog, Microsoft Copilot and ChatGPT’s agentic checkout. Because there is no product page in between, the record has to carry the whole answer itself: catalog metafields plus the agentic description Shopify hands those agents.

HOVER A RAIL — WHAT MAKES IT DIFFERENTTAP A RAIL — WHAT MAKES IT DIFFERENT
RAIL 1 ⓘ Google’s feed Merchant Center API
WHAT MAKES THIS RAIL DIFFERENT

What it buys you: every Google shopping surface at once — Shopping ads, free listings, AI Overviews and AI Mode all read this one record, and the same enrichment lifts the paid and the unpaid together. It is the only rail that carries the six conversational attributes AI Mode reads.

A structured feed pushed to Google through the Merchant Center API. Google defines every field, so a missing or malformed one quietly limits which queries you are eligible for — in Google’s own words, providing all relevant attributes “increases its chances of showing for the most relevant search queries.” Google also grades what you send: a disapproved listing drops off every surface above — the next section shows how we watch and fix those verdicts.

Shopping adsfeed attributes
Free listingsfeed attributes
Performance Maxfeed-only campaigns
Google AI Mode6 conversational attributes
RAIL 2 ⓘ The crawler rail on-page JSON-LD
WHAT MAKES THIS RAIL DIFFERENT

What it buys you: the assistants Google’s feed never reaches — ChatGPT, Perplexity, Claude, Gemini — reading your products straight off your own page. Nothing to submit, no approval queue, and it is live the moment the page updates.

The same facts are published as JSON-LD on your product page, so anything that crawls the web can read them there. The trade: it can only ever show what is genuinely on the page.

ChatGPT
Perplexity
Gemini
Claude
AI Overviews
RAIL 3 ⓘ Shopify’s agent rail catalog · MCP · ACP
WHAT MAKES THIS RAIL DIFFERENT

What it buys you: the order without the visit. A shopper can find your product and buy it inside the conversation, so you are sellable to people who never open your site.

Your catalog published into Shopify’s own agent plumbing — the Shopify Catalog and your store’s MCP endpoint — which is what feeds the Shop app, Shopify’s global catalog, Microsoft Copilot and ChatGPT’s agentic checkout. Because there is no product page in between, the record has to carry the whole answer itself: catalog metafields plus the agentic description Shopify hands those agents.

Shopify MCP agentscatalog · MCP
ChatGPTvia ACP
Microsoft Copilot
Shop appby Shopify
05 GOOGLE GRADES EVERY PRODUCT YOU SEND

Google quietly rejects listings — a missing GTIN, an incomplete attribute, a value that doesn’t match your product page — and a rejected product drops off every surface the feed serves. Most merchants find out from the sales dip. We find out from Google, the moment it happens.

LIVE MERCHANT CENTER · WHAT GOOGLE DECIDED the moment Google decides
14:02:11
Insulated Steel Bottle — 500 ml
missing GTIN[gtin]
DISAPPROVEDgoogle
APPROVEDafter fix
14:02:14
Everyday Tote — canvas
value doesn’t match the page[color]
DISAPPROVEDgoogle
APPROVEDafter fix
14:02:19
Trail Sneaker — 42 EU
missing required attribute[size]
DISAPPROVEDgoogle
APPROVEDafter fix
WATCHED
every product · live
EXPLAINED
the exact issue code
FIXED
by the same enrichment
WHY IT MATTERS
A disapproved listing is invisible at any budget.

Approval is Google’s call — we make sure your data never gives it a reason.

approval is google’s call
the data is ours
06 EVERY ORDER STAMPED WITH THE SURFACE THAT SENT IT
Orders SHOPIFY ADMIN
ORDER ATTRIBUTED SURFACE RECEIPT TOTAL
#1042 ChatGPT · agentic checkout UCP RECEIPT $34.00 #1041 Google AI Mode UCP RECEIPT $68.00 #1040 Perplexity UCP RECEIPT $34.00 #1039 Shop app UCP RECEIPT $102.00
THE STAMP ON ORDER #1042
surfacechatgpt · acp
railshopify agent rail
matched_oncapacity · care
written_toorder metafield
ATTRIBUTED
What it buys you: one number you can act on. We pull what Shopify and Google each report for these orders and reconcile it against our own receipts, so you are not left choosing between three dashboards that disagree.
07 ONE SOURCE, THREE RAILS, ALWAYS IN SYNC
RAIL 1 Google’s feed Merchant Center API
RAIL 2 The crawler rail on-page JSON-LD
RAIL 3 Shopify’s agent rail catalog · MCP · ACP

One App.
Your Products on Every Agent Surface.

What it buys you: change a product once in Shopify and every rail is updated for you. No re-export, no second catalog to maintain, and no drift between what Google sees and what an agent sees.

Install it on your store — enrichment starts the moment you connect.

Illustration — generic product, invented data
How it works

You connect. We enrich. You stay in control.

Automatic, but never out of your hands. Switch it on once and the catalog syncs, the enrichment runs and all three rails publish on their own — every value still yours to approve, edit or reject before it ships.

01You

Connect

One click connects your Shopify store and the catalog syncs itself. You decide which products go into enrichment.

One click · your Shopify store
02UCP Fluent

We enrich automatically

We build identity, attributes and Q&A, publish them to all three rails, and keep everything in sync.

Automated · three rails · in sync
03You

You approve

Review, edit, or add your own facts. Nothing we are unsure about ships until you approve it.

Your call · uncertain values wait
The mechanism

Bids decide the auction. Attributes decide entry.

  • You cannot win an auction your product was never entered into.
  • Google: providing all relevant attributes “increases its chances of showing.”
  • Performance Max and Shopping serve straight from your Merchant Center feed.
  • We enrich that feed, plus custom_label and product_type for your bidding.
  • In our ladder test, exact-product naming rose from 0% to 45% with attributes.
One feed, measuredReal sample · anonymized
Feed as exported8 field types
populated
missing attributes

The queries those missing attributes would enter are queries these products never compete in.

Enriched with UCP Fluent23 field types
+521 searchable terms that appeared nowhere in the feed

Same products, same budget — entered in more relevant queries.

Why now

The AI-referred buyer is already here.

Half of those sessions land straight on a product page — 2.5× the rate of traditional search — and they order at roughly twice the new-buyer rate of other channels. They arrive already decided, which means the deciding happened in the answer.

UCP Fluent puts your products in position to be that answer: machine-readable on all three rails, with a receipt on every attributable order so you can see exactly what the channel earned.

From our own research

Listed isn’t chosen.

Being present on agent surfaces is becoming the baseline. What decides the rest is how much of the buyer’s question your record can actually answer.

Our ladder experimentsix engines · p = 0.031
Exact product named
0%45%
Base question
0%
+1 attribute
13%
+2 attributes
38%
+3 attributes
45%

The same buying question at four lengths — each step adds one product attribute. Every engine returned 0% at step one.

1Identity

An upload-ready GS1 file.

GTINs sequenced from your prefix and GPC bricks assigned automatically. You upload, GS1 registers.

2Depth

More questions you can win.

Your products become eligible for more of the questions buyers actually ask an agent.

3Consistency

Identical everywhere.

Your product data reads the same on every surface an agent queries — feed, page and agent rail.

4Proof

Sales you can trace and trust.

A receipt on each attributable order — in your own Shopify orders and in your UCP Fluent dashboard.

One feed · two channels

Continuous enrichment, boosting organic visibility and paid channel performance.

Google: “Providing all relevant attributes for a product increases its chances of showing for the most relevant search queries.”

Bids decide whether you win the queries you’re in; attributes decide which queries you’re in. One enriched record does both jobs.

SINGLE SOURCE OF TRUTH
One enriched record
01 · ORGANIC
Named in the answers

Because the attributes are there, your product is eligible for the questions buyers actually ask.

WHERE
AI Overviews AI Mode Gemini ChatGPT Perplexity Copilot
$0 media spend
no bid, no CPC
02 · PAID
Ads running on better data

Google’s ad surfaces serve from the Merchant Center feed; Shop Campaigns reads your Shopify catalog. We enrich both — no second integration.

WHERE
Performance Max Shopping Demand Gen Shop Campaigns
1 record
feed and catalog, in sync
Run an agency?See For Agencies
// Founding Pilot

Apply for the founding pilot.

Founding pilot: free for selected merchants — full-catalog enrichment, direct Merchant Center integration, hands-on AI-commerce consulting. In exchange we publish the measured results. Applications reviewed personally, limited seats.

01

Enrichment

Full-catalog enrichment during the pilot — free for selected merchants.

02

Integration

Direct Merchant Center integration plus hands-on AI-commerce consulting from the founding team.

03

The trade

We publish the measured results as the case study.

30 minutes, your catalog, all three rails — plus the receipts you’d see in your own orders. Book a demo →

Manual review process. Zero automated spam.