Instead of one universal checkout protocol for AI shopping agents, could a stack of specialized standards work better together?
Can a single protocol work across multiple AI agents and payment providers?
This explores whether agentic commerce can run on one shared standard that any AI agent and any payment provider could use, or whether the problem is too varied for any single protocol to cover.
This explores whether one protocol could connect any AI shopping or service agent to any payment provider. The short answer from the corpus is probably not, and that may be fine. The more promising route is a stack of protocols, each handling one part of the problem, that work together. Boboev's analysis of the current crop (OpenAI's ACP, Google's AP2 and UCP, Visa's TAP, Web Bot Auth) finds they mostly don't compete. They split the work into layers: discovery (finding products), application (running the checkout), governance (who authorized what), and identity (who the agent is). Merchants add layers depending on how much fraud risk they're willing to take on Do agentic commerce protocols compete or complement each other?.
There's a deeper reason one protocol struggles, and it comes from research on agent communication generally. A survey of nine agent protocols found a trilemma: a protocol can be versatile, efficient, or portable, but not all three. Rigid formats like MCP are fast and easy to carry between systems but can't adapt to new situations. Flexible formats adapt, but agents waste effort negotiating terms with each other Can agent protocols be efficient, versatile, and portable simultaneously?. Payments sharpen this tradeoff. An IMF analysis argues that the agent's probabilistic decision-making ('should I buy this?') has to stay separate from the predictable, legally enforceable layers of authorization and settlement, because money movement can't depend on a model's best guess Should AI agents make payment decisions independently?. A protocol that tried to cover both would be either too loose for banks or too rigid for agents.
So the practical question becomes what the layers have in common. Identity keeps coming up. McKinsey points out that fraud systems were built to read human signals like device fingerprints and typing behavior, and agents don't produce those. Its proposed replacement is cryptographically signed mandates Does agent-initiated commerce break existing payment infrastructure?. Visa's Trusted Agent Protocol is one concrete version: agents sign each request so a merchant can check who the agent is, which consumer it represents, and what it intends to do Can cryptographic signatures help merchants identify trusted AI agents?. A more general academic approach extends OAuth, the system behind 'Sign in with Google', into a chain of tokens that links each agent action back to the human who authorized it, including the scope of that permission and whether it has been revoked How can AI agents prove who authorized their actions?. When this layer is missing, the result looks like Amazon blocking Meta's Muse agent: access gets settled by terms-of-service blocks, not by a standard What identity standards do shopping agents actually need?.
The corpus also has a lesson on how cross-ecosystem standards tend to win. Coordination layers succeed when they wrap and bridge existing protocols, not when they try to replace them, because each bridge pays off on its own without forcing everyone to rewrite their systems Should coordination protocols wrap existing systems or replace them?. Some interoperability is already happening outside the card networks. Chainalysis tracks x402, a crypto-based standard agents use to pay each other for services like model inference, with US wallets making up most of that spending Which countries dominate agent-to-agent AI payments on chain?.
The part you might not expect to matter is business models, which may decide this more than engineering does. Waller compares agentic commerce to the split between open-loop card networks like Visa and closed-loop ones like American Express: some platforms will let any agent in, and others will require their own Will agentic commerce platforms adopt open or closed models?. Azhar argues that an agent paid through merchant referral fees can't be loyal to both the merchant and the user, and that these incentives, not the technology, decide which platforms build agents and which block them Can an AI agent serve both merchant and user interests fairly?. A universal protocol could solve the technical handshake and still leave open the question of whose side the agent is on.
Sources 11 notes
Boboev argues ACP, AP2, UCP, TAP, and Web Bot Auth each solve different layers—application, governance, discovery, and identity—creating a capability ladder merchants stack rather than choose between, with each layer absorbing different levels of fraud risk.
A taxonomy of nine protocols reveals that rigid-schema protocols like MCP maximize efficiency and portability but sacrifice versatility, while evolving-schema protocols buy versatility at the cost of negotiation overhead. No protocol achieves all three.
An IMF analysis argues that agent-mediated payments must split probabilistic AI decision-making (Layer 1) from deterministic authorization and settlement (Layers 2–3) because core payment infrastructure requires predictability and legal enforceability that probabilistic reasoning cannot guarantee.
McKinsey argues that payment gateways and fraud engines built around explicit human identity and intent face significant strain when autonomous agents initiate transactions. The firm proposes cryptographically signed mandates as a replacement for human-observable signals fraud systems traditionally relied upon.
Visa developed Trusted Agent Protocol with Cloudflare to enable merchants to authenticate AI agents through agent-specific cryptographic signatures carrying intent, consumer recognition, and payment data. The protocol builds on HTTP Message Signatures and aims to bridge the gap between AI-assistant adoption and merchant trust.
Show all 11 sources
A three-token scheme (User ID, Agent ID, and Delegation Token) signed by human delegators creates cryptographically verifiable authorization chains. Services can now verify who authorized an agent, under what scope, and when revocation occurred.
Amazon's block of Meta's Muse reveals that AI agents need standardized ways to identify themselves and prove authorization to retailers—separate from consumer-side credential security. Without such protocols, retailers enforce access through terms-of-service blocks rather than negotiated standards.
Research shows that agent coordination standards achieve adoption by composing existing protocols like MCP and DIDComm under a shared substrate, rather than competing to replace them. Bridging lets value accrue incrementally without forcing ecosystem-wide rewrites.
Chainalysis geolocated x402 transactions show US wallets account for 88.1% of AI model inference payments and 41.8% of agent-framework payments, while Chinese wallets concentrated in security services. This reflects a shift from speculative memecoin activity toward durable agent-to-agent infrastructure spending.
Waller argues agentic commerce platforms will divide between open systems allowing multiple agents and closed systems requiring proprietary ones, mirroring payment network structure. This choice will shape market concentration, vertical integration, and standards adoption.
Azhar argues that agents like Meta's Muse, which earn referral fees from merchants like Expedia, face structural conflicts that prevent unbiased recommendations. The incentive to collect fees, not technology, determines which platforms build or block such agents.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- Payments in the Age of AI Agents
- How Agentic AI Will Reshape Payments
- The agentic commerce opportunity: How AI agents are ushering in a new era for consumers and merchants
- From Product Search to Preference Articulation: The Economics of Agentic Commerce
- Infrastructure for AI Agents
- Authenticated Delegation and Authorized AI Agents
- Magentic Marketplace: An Open-Source Environment for Studying Agentic Markets
- When Agents Shop for You: Role Coherence in AI-Mediated Markets