MatchToken binds intent, proposal, and merchant
It is not a UTM. It is the signed transaction identity agents, merchants, and settlement can agree on.
MATCHTOKEN ATTRIBUTION
Prove which AI recommendation became an order.
SellToAI does not stop at a tracked link. Each accepted proposal carries a recId and MatchToken through checkout, then Shopify/TikTok order webhooks attach the outcome back to the original intent, evidence card, merchant, and channel.
1
Intent
A buyer asks inside an AI host; SellToAI records the request context.
intent_id
2
Evidence proposal
The answer carries product proof, reason, merchant promise, and a buy path.
proposal_id
3
MatchToken
The accepted proposal gets a signed identity that survives checkout handoff.
match_token_id
4
Checkout click
The shopper leaves the AI surface through a path-bound buy link.
checkout_url
5
Order feedback
Shopify/TikTok webhooks return paid, cancelled, or refunded outcomes.
attribution_orders
What becomes measurable
Attribution is a product surface, not an internal report.
This page is the public version of the loop. The protected ops console still exists for raw dashboards, but the core value must be understandable before anyone logs in.
Shopify/TikTok orders close the loop
Order webhooks attach paid, cancelled, and refunded outcomes to the original recId and MatchToken.
Evidence can be measured against conversion
Once clicks and orders return, SellToAI can compare creator proof, community proof, risk, and merchant SLA against actual outcomes.
Every stage has an object to inspect
Intent, proposal, match session, MatchToken, click, and order are separate records, so failures are debuggable.
Closed loop
What the merchant and AI host both get back
The strategic wedge is not checkout ownership. It is making proof-backed recommendations measurable after the shopper leaves the AI surface.
Channel, intent, product fit, and merchant response live together instead of disappearing into a generic redirect.
Evidence Card fields make the recommendation auditable before the buyer clicks.
SellToAI keeps the merchant-owned checkout path and carries attribution metadata through it.
Webhook events turn clicks into paid, cancelled, refunded, or missing-order signals.
Evidence quality, merchant reliability, and channel performance become ranking inputs.
Why it matters
The winner in AI commerce will know why a recommendation worked.
Catalog access and checkout will become table stakes. The hard part is proving that a specific evidence-backed recommendation created a specific order.