ACO — Agentic Commerce Optimization

Where the term comes from

ReFiBuy coined Agentic Commerce Optimization and builds a six-stage closed-loop platform around it, per their public positioning as of mid-2026. The framing foregrounds data completeness, attribute normalization, and agent-legible product structure — the evaluation work an autonomous buyer would otherwise have to reconstruct from prose before ranking one SKU against another. eCommerce Insights defines the term neutrally: a competitor coined it, but the practice it names is real and applies to any catalog.

ACO sits closest to the agent end of the category's naming spectrum. GEO emphasizes answer citation broadly; AEO emphasizes answer engines specifically; ACO emphasizes the catalog an agent shops from. In practice the same PDP changes often serve all three.

Why it matters for ecommerce

Agentic commerce is arriving in stages. AI engines already research products before shoppers do; pre-purchase agents — ChatGPT Shopping, Perplexity's Buy with Pro — now draft carts; and agent-completed checkout via the Agentic Commerce Protocol (OpenAI/Stripe) and Google's Universal Commerce Protocol is in pilot as of mid-2026. At every stage the constraint is the same: an agent can only recommend, draft, or buy what it can read.

When an agent does the evaluating, copy written for human emotional resonance matters less; structured attributes, clean schema, and reliable price and availability data matter more. For Shopify brands ACO translates into concrete tasks: complete variant metadata, normalized attributes across a collection, typed metafields, review signal exposed via schema. The same work raises citation rates on answer surfaces today, which is why most teams run one program rather than two.

ACO readiness: an example

A Shopify brand selling ceramic mugs would measure ACO readiness by asking, per SKU, whether an agent could answer: volume in ounces, dishwasher-safe status, microwave-safe status, stoneware versus porcelain, handle style, set size. If those attributes live only in description prose, an agent may or may not extract them. If they live in typed Shopify metafields with consistent namespaces and surface in Product JSON-LD, the agent can — and the SKU survives elimination rounds it would otherwise lose. When a ChatGPT user asks for "a gift mug under $50 that is dishwasher safe," candidates missing the dishwasher attribute are dropped before price or reviews are even weighed.

ACO vs agentic commerce

The two are often conflated. Agentic commerce names the phenomenon — commerce in which agents act on the shopper's behalf. ACO names the practice of preparing a catalog to perform inside that phenomenon. A brand does not have to bet on which checkout protocol wins to start ACO work: the groundwork (schema, crawler admittance, machine-readable price, availability, and policies) is shared across ACP, UCP, and whatever follows.

How eCommerce Insights measures it

The agent-readability score is eCommerce Insights' per-SKU ACO measurement: Product JSON-LD completeness, robots.txt admittance for AI crawlers, machine-readable price and availability, discoverable returns and shipping policies, and agentic-checkout wiring where applicable. The free Agentic Readiness Grader runs the same checks on any PDP.

Related terms

• Agentic commerce — the phenomenon ACO prepares a catalog for.
• AEO (Answer Engine Optimization) — the answer-surface sibling discipline.
• GEO (Generative Engine Optimization) — the category umbrella with the largest search volume.
• Agent-readability score — the per-SKU measurement of ACO readiness.
• Product schema — the structured-data layer most ACO fixes land in.