tech
Six layers your agent has to handle. Most products have only thought about two. + a responsibility-layer audit.
For most of the internet’s history, a purchase has been a human action that everyone in the chain could see. A person clicked a button on a merchant’s site. A payment credential moved through a processor. A network or wallet weighed the risk. The merchant took responsibility for the order. The records were imperfect, but everyone agreed on the shape of the evidence: a human was present, a page was shown, a credential was used, a final action was taken.
TL;DR
- Agentic commerce involves software agents holding wallets, signing authorizations, and paying merchants directly, unlike traditional human-driven purchases.
- This new model raises questions about authorization and evidence, extending beyond the transaction itself to touch upon identity, fraud, and liability.
- The market is fragmenting into different protocol camps, each attempting to define and control various aspects of agentic commerce.
- Key players like OpenAI, Stripe, Shopify, and Google are involved in the development and contention over these new commercial protocols.
- The article highlights a need for a responsibility-layer audit to identify who owns each part of an agentic purchase.
- The future of online buying will be shaped by the resolution of these disputes over commercial trust and regulation.