Agentic commerce

What it is

Agentic commerce describes transactions mediated by an AI assistant that can take actions, not just answer questions. The agent may read product feeds, call merchant APIs, apply saved payment credentials and confirm the order inside the chat or assistant surface. Merchants participate by exposing structured product, stock and pricing data and by supporting checkout protocols that agents can call.

Why it matters

When an agent selects the shortlist, classic ranking, merchandising and on site persuasion have less influence on the outcome. Visibility depends on whether your catalogue is machine readable, accurate and trusted by the systems agents rely on. Merchants who are absent from those pipes can lose sales without ever seeing a lost session in analytics.

How it works

Practitioners start by cleaning product data: complete titles, attributes, GTINs, stock status, pricing and delivery terms in feeds and structured data. They then check eligibility for agent facing checkout routes, such as OpenAI's Instant Checkout built on the Agentic Commerce Protocol with Stripe, and equivalent programmes from other platforms. Finally they instrument referrals from assistant domains so agent driven orders can be separated from ordinary organic and paid traffic.

When it applies

It applies to any business selling products or bookable services where an assistant could plausibly complete the task, especially repeat purchases, commodities and comparison heavy categories. It matters most once a meaningful share of research starts inside an AI assistant rather than a search engine.

Examples

• A shopper asks an assistant to reorder the same coffee beans and the order is placed without visiting the retailer's site.
• A homeware retailer fixes missing size and material attributes in its feed so agents can match it to specific queries.
• A DTC brand enables an agent facing checkout route and tags the resulting orders separately in its order management system.

How it is measured

• Orders and revenue attributed to AI assistant referrers or agent checkout routes
• Product feed completeness and error rate across required attributes
• Rate at which your products appear in assistant generated shortlists for tracked prompts
• Return and cancellation rate on agent sourced orders compared with site orders