Sign in

Alipay Just Made the Agentic Commerce Protocol Stack 4-Deep

7 min read
Share:
Large emerald-gradient hero typography reading 4 PROTOCOLS LIVE on slate canvas with ALIPAY and ANT GROUP wordmarks, no bottom banner.

Ant Group launched China's first agentic commerce protocol this week. The agentic commerce protocol stack now has four named entries. If you sell into US, EU, and Asia, you do not get to pick one and ignore the rest. Your catalog has to read cleanly to all of them.

TL;DR: Alipay debuted AI Wallet, Token Pay, and the Agentic Commerce Trust Protocol (ACTP) on May 26, 2026, per PYMNTS. That makes four named agentic commerce protocols live in market alongside UCP (Google/Shopify), ACP (OpenAI), and AdCP (Scope3/Yahoo). Brands selling cross-border now need a single source of product data that publishes cleanly to every protocol, because picking one and ignoring the rest leaves regional revenue on the table.

What Alipay Actually Launched

Ant Group announced three pieces of agentic payment infrastructure on May 26, 2026, all positioned as one stack. The AI Wallet is searchable inside the Alipay app today, per Finews.Asia. Token Pay is a new payment instrument built specifically for agent-driven transactions. The Agentic Commerce Trust Protocol (ACTP) is the common-language framework that lets AI systems and service platforms collaborate on payments and identity.

Coverage cascaded across Asia trade press through May 27 and 28. TechWire Asia framed AI Wallet as "a control layer sitting between the user and the agent," which is the same framing US payment networks have used for Visa Trusted Agent Protocol and Mastercard Agent Pay. The architecture rhymes. The geography does not.

ACTP is being positioned as "China's first" agentic commerce trust protocol. That signals two things at once. Ant Group expects more Chinese protocols to follow, and Ant wants to be the one that other Chinese platforms reference instead of importing UCP or ACP wholesale.

The Protocol Stack Just Went 4-Deep

As of May 29, 2026, there are four named agentic commerce protocols live or in active deployment:

Protocol Sponsor Surface Primary geo
UCP (Universal Commerce Protocol) Google + Shopify AI Mode, Gemini, Merchant Center US / global
ACP (Agentic Commerce Protocol) OpenAI ChatGPT shopping US / global
AdCP (Ad Context Protocol) Scope3 + Yahoo Ad-side agent buying US
ACTP (Agentic Commerce Trust Protocol) Ant Group / Alipay Alipay app + Token Pay China / APAC

UCP has 26 confirmed retail deployments according to our recent reporting on Google's "Beyond the Catalog" event. ACP underpins ChatGPT's shopping results and the OpenAI product feed spec. AdCP went live earlier this spring with Scope3 and Yahoo announcing it as the ad-side parallel. ACTP is now the China-specific entry on the same map.

A retailer selling in the US, EU, and APAC at meaningful volume in 2026 has to be readable by four protocols, not one. That is the new floor.

That is a structural change, not a press release. Each protocol expects subtly different attribute coverage, identity conventions, and trust signals. A product feed that publishes cleanly to Google Merchant Center does not automatically publish cleanly to Alipay's ACTP. And the spread is widening as Visa, Mastercard, American Express, and FIDO Alliance each layer their own identity and trust frameworks on top, as PYMNTS reported on May 28.

Why Mid-Market Brands Should Care

There is a tempting read here that says ACTP only matters if you sell into China. That read is incomplete.

First, the protocol-multiplication pattern is real. Six months ago, the conversation was UCP versus ACP. Today it is four protocols and counting. Visa's Trusted Agent Protocol and Mastercard Agent Pay are layering on top, and American Express ACE is doing the same on the card-network side. The brands that win 2026 to 2027 will not be the ones who picked the right protocol. They will be the ones who built a catalog that publishes cleanly to whichever protocol an agent is using when it asks.

Second, the analyst consensus is finally lining up. Forrester published "The State of Agentic Commerce, in Mid-2026" on May 28 with a formal Agentic Commerce Framework that opens with "evaluate your agentic fit." Forrester is saying out loud what mid-market merchandising leaders have been telling each other privately for months: this is not a build-it-once problem, it is an audit-and-publish-continuously problem.

Third, the readiness gap is documented. Digital Transactions reported on May 28 that 73% of online merchants are not ready for agentic commerce. That number is from a third-party survey, but it lines up with what every brand audit we have run shows. The catalogs are not ready, and protocols are arriving faster than catalogs are being fixed.

The Control-Layer Framing Is the Real Tell

Read the Alipay announcement closely and the phrase "control layer between the user and the agent" appears in every secondary write-up, including Yahoo Finance. That is not marketing accident. It is the same language Visa used to describe Trusted Agent Protocol. It is the same language Forrester used in its Q2 framework. It is the same language a CIO uses in a brand-side audit conversation.

The control-layer framing exists for one reason: nobody, on either side of an agentic transaction, fully trusts the agent. The consumer wants a layer that lets them see what the agent intends to buy. The brand wants a layer that lets it see what agents are pulling from its catalog and how they are presenting it. The payment network wants a layer that lets it verify the agent's mandate before settling the transaction.

For consumers, Alipay is shipping that control layer as an AI Wallet inside the Alipay app. For brands, the equivalent control layer is the multi-channel AI commerce audit stack: knowing what AI agents see when they query your catalog, which surfaces they cite you on, and where your product cards are getting clipped or skipped. The product data is the lever. The visibility into how that data renders across surfaces is the control layer.

Where Paz Fits

Paz.ai, an agentic commerce optimization platform, ships the brand-side control layer for AI shopping surfaces. The platform monitors how a brand appears across ChatGPT, Gemini, Perplexity, Microsoft Copilot, and Amazon's AI shopping assistant, then publishes corrected product data back to the surfaces that consume it. Protocol-neutral is the whole point. A catalog that is ACP-clean, UCP-clean, AdCP-clean, and ready for ACTP-style requirements is not a four-week sprint per protocol. It is one structured catalog that publishes correctly to each.

The Q2 2026 protocol map is genuinely new. The Q1 map had two named entries and a lot of hand-waving about which one would win. The Q2 map has four entries and no winner emerging. That is the environment brands are actually selling into, and it is the reason the AI commerce protocols layer reads as infrastructure rather than experiment in 2026.

What to Do This Week

  1. Inventory your protocol exposure. List every market you sell into and which agentic commerce surfaces (ChatGPT, Gemini, Alipay AI Wallet, Amazon Rufus) are live in each. If you sell into China at any volume, ACTP is not optional for your roadmap. If you sell only US/EU, you are still on the hook for UCP, ACP, and AdCP.
  2. Audit your catalog against the four protocols, not one. Run your top 100 SKUs through each protocol's required attribute set. Most catalogs that pass Google Merchant Center fail ChatGPT Product Feed Format on identity attributes and miss product data enrichment requirements that ACTP-style trust protocols are likely to require.
  3. Read Forrester's Q2 framework. The "evaluate your agentic fit" step in Forrester's State of Agentic Commerce is the cleanest assessment framework any top-3 analyst has put out this year. Use it to sequence your own internal readiness review.
  4. Pick the control layer. Decide who, internally, owns visibility into how AI agents see your catalog. If the answer is "nobody," that is the gap. Forrester just called it out by name.
  5. Stop optimizing for the protocol of the month. UCP was the protocol of February. ACP was the protocol of March. ACTP is the protocol of this week. The structural answer is a catalog that is readable by all of them, not a sprint for each.

Frequently Asked Questions

What is the Agentic Commerce Trust Protocol (ACTP)?
ACTP is the protocol Ant Group / Alipay announced on May 26, 2026, as China's first agentic commerce trust protocol. It is the common-language framework that standardizes how AI systems and service platforms collaborate on agent-driven transactions, paired with Alipay's AI Wallet and Token Pay payment instrument inside the Alipay app.

How does ACTP differ from UCP and ACP?
UCP (Google/Shopify) is product data and catalog focused. ACP (OpenAI) is product feed and shopping-result focused. ACTP is trust and payment focused, sitting closer to the agent mandate and identity verification layer than to the catalog layer. They overlap, but each emphasizes a different slice of the agentic commerce transaction.

Do US brands need to care about ACTP?
Indirectly, yes. ACTP is the fourth named protocol on the agentic commerce map, and protocol multiplication is the pattern that matters. US brands selling into APAC do need ACTP readiness on the roadmap. US-only brands should treat ACTP as evidence that the protocol stack is widening, not consolidating.

Is Forrester saying agentic commerce is hype?
Forrester's Q2 2026 take is more nuanced than "hype." The blog says "hype is running ahead of behavior," meaning consumer trust and adoption lag the headlines, while answer-engine discovery is converting well even when checkout happens off-platform. The recommendation is to "evaluate your agentic fit" before investing in agent storefronts.

What is the "control layer" framing in agentic commerce?
The control layer is the visibility and intervention surface that sits between a consumer (or brand) and an autonomous AI agent. Alipay's AI Wallet is the consumer-side control layer. Brand-side, the equivalent is an AI readiness score and continuous monitoring of how a catalog renders across surfaces.

Will the protocol stack consolidate?
Probably not in 2026. Each protocol sponsor has structural reasons to keep its own stack (geo regulation, platform economics, vertical specialization). Brands should plan for protocol multiplication through at least 2027, not consolidation.

The protocol stack got bigger this week. Catalogs did not. That is the gap to close, and Forrester just put a name on the assessment step that closes it.

How AI-ready are your products?

Check how ChatGPT, Google AI, and Perplexity evaluate any product page. Free score in 30 seconds.

Run Free Report →

Related posts

AWS Just Started Selling Its Agentic Shopping Stack. It Doesn't Fix Your AI Visibility.

AWS Just Started Selling Its Agentic Shopping Stack. It Doesn't Fix Your AI Visibility.

TL;DR: AWS launched the Agentic Shopping Assistant (ASA) on May 27, 2026, packaging the same Claude Haiku 4.5 architecture that powers Amazon's own Alexa for Shopping for outside retailers. Tapestry's Kate Spade is live with a Gift Concierge after a 2.5-month build; deployment runs as little as 60 days. First time a top-three hyperscaler sells agentic shopping infrastructure as a packaged vendor product, and it changes the buyer conversation. But ASA puts an Amazon-powered agent inside yo

May 28, 2026
min read
Read article
Google's UCP Has 26 Deployments. AI Mode Has 1 Billion Users.

Google's UCP Has 26 Deployments. AI Mode Has 1 Billion Users.

Four months after Google announced the Universal Commerce Protocol at NRF with twenty co-developers and twenty-something more endorsers, an independent dashboard found 26 public deployments across three million scanned domains. None of them are Walmart, Target, Etsy, Wayfair, or Shopify, the brands Google put on stage. TL;DR: An Originality.ai dashboard launched May 21, 2026, scanned 3M+ websites for the UCP manifest and found 26 implementations. Zero of the named co-developers have a p

May 27, 2026
min read
Read article
Salesforce Just Put $5T on Agentic Commerce. Most Brands Will Miss It.

Salesforce Just Put $5T on Agentic Commerce. Most Brands Will Miss It.

Salesforce and Publicis Sapient just released a joint thesis pegging the global agentic commerce opportunity at $3 trillion to $5 trillion by 2030. In the same report, they cite an MIT NANDA finding that 95% of generative AI initiatives fail to deliver measurable business impact. Those numbers are not in tension. They are the same story. TL;DR: The biggest dollar figure yet on agentic commerce just dropped from Salesforce and Publicis Sapient: $3-5T by 2030, with a two-to-three year win

May 26, 2026
min read
Read article