AI agent autonomous commerce workflow — UCP protocol visualization

Will’s Take: UCP’s Merchant of Record Promise Is a Checkout Interface Play — Not a Power Position

Payment & Settlement Standard: November 10, 2026 • UCP-PAY-SPEC v2026-11-10

Production Validated

Every article written about the Universal Commerce Protocol includes some version of the same reassuring sentence: merchants remain the merchant of record. Google says it. Shopify says it. The partners say it. The analysts quoting the partners say it. It appears in official documentation, press releases, and every implementation guide published since January 2026.

Nobody is asking what that phrase actually means — or more importantly, what it doesn’t mean.

I want to be clear upfront: I’m not arguing against UCP adoption. The protocol solves a real problem, and sitting it out likely costs more than participating. But the gap between what “merchant of record” implies and what it delivers operationally is wide enough to drive a commerce strategy through — and if you don’t see the gap, you’ll build your business around assumptions that aren’t true.

The Phrase Everyone Repeats Without Examining

PAYMENT SPECIFICATION MATRIX: Will’s Take: UCP’s Merchant of Record Promise Is a Checkout Interface Play — Not a Power Position (SPEC v2026-11-10)
Commerce Dynamic Merchant of Record (MoR) Aggregator Open UCP Direct Merchant Model Power Distribution Take-Rate Reality Strategic Outcome
Customer Data Ownership Aggregator owns buyer relationship & data Merchant owns 100% customer record Merchant retains sovereign data Aggregator takes 5-10% fee First-party data preserved
Tax & Nexus Liability Aggregator assumes global tax remittance Automated edge tax engines remit for merchant Merchant retains tax control MoR charges premium margin Zero tax mark-up fees
Dispute & Refund Control Arbitrary MoR chargeback deductions Direct merchant customer resolution Merchant controls policy Aggregator charges $25/dispute Fair dispute handling
Payment Gateway Choice Locked to MoR proprietary processor Merchant selects best-fit acquirers Free market competition MoR pads interchange fees Lowest wholesale rates
Checkout UX Customization Rigid generic MoR hosted checkout Native in-assistant conversational buy Tailored brand experience MoR imposes platform branding 100% brand loyalty
Platform Governance Centralized corporate terms of service Decentralized open standards consortium Democratic industry consensus MoR can deplatform merchants Immune to deplatforming

“Merchant of record” is a legal and financial designation. It means you are the entity responsible for the transaction. You handle taxes. You handle returns. You handle chargebacks. You carry the liability. That is what the phrase means, and under UCP, it is accurate — you do carry all of that.

What it does not mean is that you control the customer relationship. It does not mean you control discovery. It does not mean you own the pre-purchase journey. It does not mean you see the full data picture of how a buyer arrived at your product, what they compared it against, or what the AI agent recommended before settling on you.

“Merchant of record” sounds like a position of power. In practice, under native UCP checkout, it is a position of responsibility — with the power sitting elsewhere.

Google Has Run This Play Before

In 2015, Google introduced Accelerated Mobile Pages. The pitch was familiar: an open standard, better for everyone, publishers keep their content. What actually happened was that Google moved the reading experience to Google’s servers, reduced publisher brand visibility to a small gray URL line, and made Google the de facto interface between readers and the publications they used to visit directly.

Publishers kept their content. They lost their audience relationship.

AMP was eventually walked back — Google acknowledged in 2021 that it would no longer require AMP for top stories eligibility. But the structural damage was done. An entire generation of mobile web publishing was shaped by the assumption that Google’s interface was the only viable path to readers.

UCP follows the same structural logic. Move the transaction to Google’s surface. Let the merchant handle fulfillment and liability. Call it open. The difference is that commerce has higher stakes than content — and the checkout is harder to walk back than a page format.

What Google Actually Controls Under UCP

Let’s walk through a native UCP transaction and note who controls each stage:

Discovery: The customer asks Google’s AI Mode or Gemini for a product recommendation. Google’s models decide which products to surface, in what order, with what framing. Google controls this.

Comparison: The AI agent evaluates multiple merchants’ structured data — pricing, availability, shipping speed, return policies. The agent decides which merchant to recommend. Google controls this.

Checkout surface: Under native checkout, the transaction happens on Google’s interface. The customer never visits the merchant’s website. Google controls this.

Payment rails: Payment processes through Google Pay via Google Wallet. PayPal support is coming, but the payment infrastructure runs through Google’s credential system. Google controls this.

Post-purchase communication: Tracking, returns, and support can flow through the same AI conversational interface. Google hosts this.

What the merchant controls: Pricing. Inventory accuracy. Fulfillment. Customer service after delivery. Tax liability. Chargeback exposure.

Read that list again. The merchant controls the cost centers. Google controls the revenue levers.

The Data You Keep vs. the Data You Don’t

Google’s documentation states that merchants retain ownership of customer data and relationships. This is technically accurate at the transaction level — you get the order, the shipping address, the payment confirmation.

What you don’t get is the intelligence layer above the transaction:

You don’t see what query the customer used. You don’t see what competitors the agent evaluated before choosing you. You don’t see the agent’s reasoning for why your product was selected — or nearly wasn’t. You don’t see cross-merchant browsing patterns. You don’t see how the agent weighted price versus reviews versus shipping speed in its recommendation.

In traditional e-commerce, a customer visiting your site generates behavioral data at every step — what they searched, what they clicked, how long they lingered, what they abandoned. That data powers your merchandising, your email sequences, your retargeting, your product development. Under native UCP checkout through AI Mode, that entire behavioral layer belongs to Google. You receive an order. Google receives an understanding of your customer.

This is the asymmetry that the phrase “merchant of record” obscures. You own the transaction record. Google owns the decision record.

The Embedded Option Exists — But Read the Fine Print

To be fair, UCP offers an embedded checkout option where the customer is redirected to the merchant’s own checkout flow. This preserves more brand control and data capture. Google’s developer documentation describes it as an option for merchants with “highly bespoke branding or complex checkout flows.”

But the incentive structure points the other direction. Native checkout is frictionless — no redirect, no new page load, no form to fill. In a world where AI agents optimize for speed and conversion, the agent will prefer the path with fewer steps. Merchants who choose embedded checkout may find their products deprioritized by agents that favor faster transaction completion.

The option exists. The ecosystem rewards the other choice.

Why I’m Still Watching — Not Warning

None of this means UCP is a bad protocol or that merchants should avoid it. The integration simplification is real. The access to AI-native shopping surfaces is valuable. The open-source specification is genuinely available on GitHub. And the alternative — being invisible to agentic commerce entirely — is worse than participating with open eyes.

What I am saying is that “merchant of record” is not a synonym for “merchant in control.” It is a liability designation wrapped in language that sounds like a power position. Every merchant adopting UCP should do so with a parallel strategy:

Build and maintain your owned checkout experience. Continue investing in direct customer relationships — email lists, SMS, loyalty programs — that don’t depend on any agent or protocol. Treat UCP as a distribution channel, not your storefront. And watch the data-sharing arrangements closely as they evolve, because Google has not fully disclosed the reporting and attribution merchants will receive for AI Mode transactions.

The merchants who thrive under UCP will be the ones who see it clearly: a powerful distribution channel with a significant power asymmetry, operated by a company with a documented history of consolidating interface control while distributing liability.

Adopt it. Just don’t confuse carrying the risk with holding the reins.

Frequently Asked Questions: Payment Rails & Settlement Standards

Q1: What is the difference between a traditional Merchant of Record (MoR) and UCP?

Traditional MoRs (like Paddle or FastSpring) act as legal intermediaries, taking a 5-10% cut while owning customer data. UCP provides an open protocol that lets merchants maintain direct customer relationships while automating tax and compliance programmatically.

Q2: Why do platforms attempt to position themselves as Merchants of Record in AI commerce?

Controlling the Merchant of Record position allows platforms to capture high processing margins and dictate terms to merchants. UCP prevents this power grab by standardizing direct protocol checkout.

Q3: How does UCP handle international tax compliance without an MoR middleman?

UCP integrates with edge tax automation services (such as Avalara, TaxJar, or Vertex) to calculate, collect, and report international VAT and sales tax directly on the merchant’s behalf.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *