For thirty years, buying online has followed the same shape: you open a website, search or browse, compare, add to cart, check out. Agentic commerce breaks that pattern. Instead of navigating a store yourself, you delegate the task to an AI assistant that searches, compares and — increasingly — buys on your behalf, inside a single conversation.
This article explains what agentic commerce is, how it works, why it's emerging now, and what it means for merchants.
What agentic commerce means
Agentic commerce is commerce mediated by an AI agent acting on behalf of a shopper. Rather than clicking through a website, the shopper expresses an intent — "find me running shoes under €100 that suit flat feet" — and the agent handles discovery, comparison, and often the purchase itself.
The agent is not a search box. It interprets a natural-language request, queries product catalogs it can read, weighs options against the stated constraints, presents a short list with real prices and availability, and, when enabled, completes the transaction and payment without the shopper ever visiting a storefront.
The shift is from shopper-navigates-store to agent-serves-shopper. The store still fulfils the order and owns the customer relationship — but the moment of discovery and decision moves into the assistant.
How agentic commerce works
Three ingredients must come together for an AI agent to complete a purchase.
A machine-readable catalog. The agent can only recommend products it can read. That means a structured product feed — titles, descriptions, attributes, prices, availability, images — supplied in a format the agent's platform understands. Without it, your products are invisible to the agent, no matter how good they are.
A discovery and matching layer. The agent interprets the request and matches it against available catalogs, ranking results by relevance to the shopper's intent rather than by ad spend. Relevance is judged on the quality and completeness of the product data.
A secure payment and checkout path. To buy inside the conversation, the agent needs a trusted way to pass payment to the merchant, with fraud protection, buyer protection, and dispute handling. This is where payment providers and emerging open standards come in.
When these three work together, a request becomes a purchase in a few conversational turns.
Why agentic commerce is emerging now
Several trends have converged.
Consumers already use AI assistants to research purchases — comparing options, checking specs, narrowing choices — and the natural next step is to let the assistant act, not just advise. At the same time, the major AI platforms have moved from answering questions to completing tasks: ChatGPT, Gemini and Google's AI Mode have introduced shopping and checkout experiences directly in their interfaces.
The missing piece — a standard way for agents to transact with merchants — is now being built. Open protocols for agent-initiated checkout and agent payments emerged through 2025 and into 2026, backed by the largest platforms and payment networks. With the rails being laid, the barrier to agentic purchases is falling fast.
The demand signal is already measurable. PayPal's Agentic Commerce Pulse Report (498 US merchants, February–March 2026) found that nearly all merchants report at least partial visibility into AI-agent traffic, while only about one in five have 80% or more of their catalog available as structured, machine-readable data. The audience is arriving before most stores are ready for it.
What agentic commerce changes for merchants
Agentic commerce reshapes several assumptions that have held for a decade.
Discovery is won on data, not layout. There is no homepage, no hero banner, no carefully designed category page in the agent's decision. What the agent sees is your structured data. Product titles, attributes, images and freshness become the storefront.
Relevance replaces placement. Agents surface products that best match the shopper's stated need. A complete, well-structured listing for the right query can outperform a bigger ad budget attached to a thin one.
The catalog becomes the channel. The same product feed that powers Google Shopping, Performance Max and Meta is now the input for a new class of buyer. Merchants who maintain a clean, enriched, continuously updated feed can extend into agentic surfaces with little extra work. Those who don't will be absent from them.
Trust and fulfilment still matter. The agent brokers the discovery and the payment, but the merchant still fulfils the order, owns the brand and handles service. Accurate availability and pricing are not just compliance details — they determine whether the agent trusts your catalog enough to surface it.
How to get ready
You don't need to predict which assistant or protocol will win to prepare — because they all depend on the same foundation: a machine-readable, high-quality, continuously updated product catalog.
Practical priorities:
- Structure and enrich the feed: descriptive titles, complete attributes, compliant images, correct categories.
- Keep it fresh: prices and availability synced frequently, so agents never see stale data.
- Distribute it where agents look: through the services and platforms that connect catalogs to AI assistants.
These are the same disciplines that already drive performance in Shopping and paid social — which is why merchants with a strong feed operation are best placed for what comes next.
The bottom line
Agentic commerce moves the moment of discovery from your website into the AI assistant. It rewards structured, accurate, up-to-date product data and makes the catalog the channel. It's early, and the standards are still settling — but the prerequisite is already clear, and it's one merchants can act on today: make your catalog readable and trustworthy for the agents that are already shopping on your customers' behalf.
→ PayPal Agentic Commerce: what it is, how it works, and what merchants need to know
→ How to make your product catalog discoverable by AI assistants
→ Discover Product Feed Marketing suite by Highstreet.io
→ Request a demo