What is Agentic Commerce?

Home > Glossary > Agentic Commerce

Agentic commerce is a commerce model in which AI agents can discover products, evaluate options, and perform commercial actions on behalf of buyers or merchants. Unlike conversational commerce, it involves delegated actions such as quoting, ordering, checkout, or post-purchase coordination.

The important shift is from assistance to execution. A conventional chatbot can recommend a product or explain a policy; an agentic commerce system can move the transaction forward by calling commerce tools, checking live inventory, creating a cart, requesting a quote, or submitting an order. That requires the agent to work with authoritative systems rather than rely only on text generated by a language model.

A production-ready agentic commerce setup usually needs:
• Machine-readable catalog, price, inventory, shipping, and policy data that an agent can query reliably.
• APIs or protocols for discovery, cart, checkout, order, and post-purchase actions.
• Identity, permissions, spending limits, and explicit confirmation rules that define what the agent may do.
• Logs, error handling, and human escalation for transactions that are ambiguous, high-value, or outside policy.

For retailers and B2B sellers, the opportunity is fewer manual steps between intent and purchase. The risk is that a confident agent can act on stale data or exceed delegated authority if the commerce stack is not designed for machine customers. Agentic commerce therefore depends as much on integration, governance, and data quality as on the AI model itself.

Example: A buyer asks an assistant to reorder approved maintenance supplies under a defined budget. The agent checks account pricing and stock, builds the order, and submits it only after the required approval or confirmation is satisfied.