Agentic commerce is commerce where an AI agent participates in product discovery, comparison, negotiation, checkout, or post-purchase work on behalf of a user or organization. The agent may recommend options, fill forms, call merchant APIs, track constraints, or initiate a purchase after receiving authorization.
The term covers a spectrum of autonomy:
- assisted shopping, where the agent researches and drafts actions;
- delegated checkout, where the user approves a specific transaction;
- recurring procurement, where the agent acts within predefined limits; and
- autonomous purchase attempts for time-sensitive scenarios such as inventory drops or ticket releases.
Agentic commerce requires more than a chat interface. Merchants need machine-readable product data, structured availability and pricing, clear return policies, authenticated APIs, fraud controls, and payment protocols such as Agent Payments Protocol (AP2).
The main technical challenge is authority. A merchant must know what the user allowed the agent to do, whether the item and price match that authorization, and who is accountable if the transaction is wrong. Verifiable intent is one approach to preserving that evidence.
Agentic commerce also changes discovery. Product pages, feeds, APIs, llms.txt, and agents.txt may become part of how agents decide which merchants are usable.
Google frames AP2 as infrastructure for AI commerce in its Agent Payments Protocol announcement.
The LLM Knowledge Base is a collection of bite-sized explanations for commonly used terms and abbreviations related to Large Language Models and Generative AI.
It's an educational resource that helps you stay up-to-date with the latest developments in AI research and its applications.
Previous