Kachyng maps the missing layer for AI agent commerce

Kachyng released a new whitepaper arguing that current agent-commerce protocols can connect AI to software and payments, but still leave enterprise governance unresolved. The paper positions Kachyng’s A-Commerce stack as the operating layer needed for identity, authority, compliance and controlled execution.

Why it matters: - Kachyng says the next phase of AI commerce is not just about finding and paying for goods. It is about whether autonomous agents can execute business actions safely inside enterprise controls. - The whitepaper argues that retail-style shopping flows are easier to automate than enterprise procurement, where approvals, contracts, entitlements and payment rules matter. - The company frames this as a missing infrastructure layer for governed AI action across traditional financial rails.

What happened: - Kachyng Inc. released "When AI Agents Go Shopping," the latest paper in its State of Agent Commerce series. - The paper was published in San Francisco on July 14, 2026. - The whitepaper maps the protocols shaping agent commerce, including MCP, A2A, UCP, ACP, AP2, Visa's TAP, Mastercard Agent Pay, x402, OAuth, OIDC and RFC 8693 token exchange. - Kachyng says the paper groups those protocols into a single layered architecture instead of treating them as separate tools. - The company made the paper available for download here.

The details: - The paper says current protocols help AI agents reach software, interact with merchants, present evidence of consent, establish network trust and initiate payment. - The paper says those protocols do not convert raw APIs, embedded business rules, contracts, approvals, entitlements and internal controls into a governed set of actions that autonomous agents can safely execute. - Kachyng draws a distinction between consumer and enterprise commerce. - Consumer commerce usually runs from search to catalog, cart, checkout and payment. - Enterprise commerce usually runs through approved suppliers, negotiated contracts, pricing agreements, requisitions, approval chains, purchase orders, invoices and settlement. - Kachyng says many current agent-commerce protocols still reflect the retail model. - The company says enterprise commerce, which it views as the larger opportunity, remains unresolved. - The paper also examines the shift from probabilistic AI reasoning to governed execution. - The paper says an AI model may propose an action, but a business system still has to decide whether to execute, stop, escalate or request more information. - The paper states: "Inference determinism is an engineering property. Decision determinism is a governance property." - Kachyng’s A-Commerce platform is built around four pillars: IDX, AGX, KYA and PRX. - IDX, or Identity Exchange, establishes who the agent is, who it represents and what authority has been delegated. - AGX, or Agent Gateway Exchange, turns raw software functions and buried business rules into governed business actions that autonomous agents can select and execute. - KYA, or Know Your Agent, extends compliance and trust controls to non-human actors before they can transact. - PRX, or Processor Exchange, connects authorized agents to existing payment processors and financial rails. - Kachyng says those four layers provide the trust, identity, authorization, compliance and payment-routing infrastructure required for A-Commerce. - Wallaja said AGX does not make the AI agent deterministic. - Wallaja said AGX makes the business outcome governed, reproducible and enforceable. - The paper’s operating model centers on two principles: "Rules, not feeds" and "Expose decisions. Enforce rules."

Between the lines: - The whitepaper is a positioning move as much as a technical map. - Kachyng is arguing that protocol-level progress is not enough unless enterprises can also govern what agents are allowed to do. - That framing places identity, authorization and policy enforcement ahead of pure model capability.

What’s next: - Kachyng is likely to continue building and marketing its A-Commerce platform around IDX, AGX, KYA and PRX. - The company is betting that enterprise buyers will want a control layer that sits between AI reasoning and transaction execution. - The broader agent-commerce market will likely keep evolving around competing standards, but Kachyng is defining the enterprise problem as governed execution rather than simple payment initiation.

The bottom line: - Kachyng’s new whitepaper says the hard part of AI commerce is not teaching agents to shop. It is making business systems decide, with governance, what those agents are actually allowed to do.