Skip to content
UCP
Menu

Satellite

Use cases

Concrete scenarios where an agent meets a merchant. For each: what the agent does, what it reads, what the merchant has to expose.

Updated : April 2026 · Primary query : agentic commerce use cases

The Universal Commerce Protocol is abstract until you stand it next to a real buying task. Below are eight concrete use cases where agents are already operating in commerce. Each one lists what the agent does, what it reads from the merchant, and what the merchant should therefore expose.

1. Conversational product discovery

Example. A user asks ChatGPT or Perplexity "best noise-cancelling headphones under €300 with hearing-aid mode".

  • Agent reads: product titles, technical attributes (ANC, transparency, hearing assist), price, availability, brand trust.
  • Merchant exposes: attribute-typed PDPs, JSON-LD Product/Offer, feed entries with full specs, structured returns.

2. Cross-merchant comparison

Example. A user asks the agent to compare three rain jackets across Patagonia, Arc'teryx and a DTC brand, under €200, shipping to France by Friday.

  • Agent reads: GTINs for dedup, technical specs, shipping windows per region, returns.
  • Merchant exposes: GTIN consistency, shipping rate settings with region granularity, price parity between PDP and feed.

3. Procurement and B2B restock

Example. A restaurant group's agent auto-restocks disposable gloves every Tuesday within a €500 budget.

  • Agent reads: bulk price tiers, SKU-level availability, B2B account endpoints, invoice and VAT behaviour.
  • Merchant exposes: B2B catalog API, tiered pricing, stable SKUs, invoice API, agent-scoped credentials.

4. Gifting by constraint

Example. A user says "find a birthday gift under €80 for a fan of Japanese stationery, ship to Berlin by Saturday, no plastic packaging".

  • Agent reads: categorization, sustainability attributes, packaging metadata, express shipping eligibility.
  • Merchant exposes: enriched taxonomy placement, material/packaging attributes, shipping cutoff times.

5. Price monitoring and trigger-buy

Example. An agent watches the price of a specific sneaker SKU and buys when it drops below €120.

  • Agent reads: price history, current price freshness, availability, agent-pay acceptance.
  • Merchant exposes: accurate price_effective_until fields, real-time feeds, webhooks for price changes, support for agent payment credentials.

6. Subscription renewal and substitution

Example. A household's agent renews a vitamin subscription and substitutes when the preferred SKU is back-ordered.

  • Agent reads: subscription terms, substitution eligibility, stock state, expected ship date.
  • Merchant exposes: structured subscription metadata, substitution policy, inventory API with ETA fields.

7. After-sales — returns and warranty

Example. An agent initiates a warranty claim for a broken blender on behalf of a user, providing the proof of purchase.

  • Agent reads: warranty terms, claim endpoint, order history, identity of original buyer.
  • Merchant exposes: structured warranty fields, returns/claim API, agent authentication on behalf of user.

8. Agent-augmented merchandiser (merchant-side)

Example. A merchant's own agent (Shopify Sidekick-style) audits its catalog, proposes attribute enrichment, and flags missing GTINs.

  • Agent reads: internal PIM, feed diagnostics, competitor public catalogs.
  • Merchant exposes: (to its own agent) authenticated PIM access, write endpoints, audit logs.

Cross-cutting lessons

Across the eight:

  • Structured policies are read in every scenario — not a nice-to-have.
  • Price and stock freshness are the two most frequent failure points.
  • Agent authentication — both agent-as-agent and agent-on-behalf-of-user — is the sharp edge.
  • B2B scenarios are technically further along than consumer because procurement APIs already exist.

Translate these into concrete steps via the readiness checklist and best practices.