The complete guide to Agentic Commerce Optimization (ACO).

ACO in one sentence

Agentic Commerce Optimization (ACO) was coined by . The practice of preparing product catalogs so AI shopping agents (including autonomous purchasing agents) can evaluate, compare, and recommend them. ACO emphasizes SKU-level product-data readiness over marketing-content optimization. It overlaps heavily with AEO and GEO when applied to ecommerce.

That is the canonical definition eCommerce Insights publishes in the glossary, and it is the definition this guide references throughout. The wording matters. ACO is about preparing catalogs, not writing content. Its unit of work is the SKU, the variant, the metafield, the availability flag. Its target is an agent that reads structured data, not a human reading marketing copy.

eCommerce Insights takes no position on 's tooling or pricing. The term ACO is useful whether or not a brand ever evaluates 's platform. Crediting the originator is a practice eCommerce Insights commits to across the category; see the taxonomy note on the homepage and the for a broader read on how the two companies differ.

Who coined ACO and why

introduced Agentic Commerce Optimization in its public materials during 2024 and 2025. The framing reflected a specific bet: that AI shopping agents — programs that evaluate and sometimes purchase on a user's behalf — would become a meaningful sales surface in their own right, requiring optimization work distinct from classical SEO, GEO, and AEO.

The bet was defensible. By Q1 2026, ChatGPT's Operator, Perplexity's Buy with Pro, and Gemini's shopping flows had all moved from research preview to limited release, signaling that the agent-mediated purchase path was no longer hypothetical. The practitioner conversation around catalog readiness followed. ACO gave that work a name.

Competitors have responded differently. uses the term Agent Engine Optimization (a reading of AEO) that overlaps substantially with ACO. Profound, and the brand-tracking vendors largely do not use either term and focus on brand-mention measurement. eCommerce Insights uses ACO as the shared category reference for catalog-data readiness and links to the full AEO ambiguity discussion for teams that need to disambiguate in practitioner conversations.

ACO names a shift most Shopify brands have not yet priced: optimizing for buyers who are not human.

How ACO differs from GEO and AEO

GEO and AEO optimize for human-facing generative surfaces — a user reads an AI-generated answer and decides. ACO optimizes for agent-facing surfaces — an AI shopping agent reads structured catalog data and decides on behalf of the user. That shift in audience has four practical consequences.

1. The unit of optimization

GEO and AEO target pages and passages. ACO targets product records — SKUs, variants, metafields, offers blocks. A well-optimized product record is richer than what a human reader needs.

2. The freshness requirement

Humans tolerate stale content. Agents do not. An agent evaluating three products against a buyer's budget needs current prices and real availability. ACO elevates the freshness of price, inventory, and fulfillment signals to a first-class concern.

3. The comparability surface

GEO and AEO optimize for citation. ACO optimizes for comparison. An agent compares Product A against Product B on weight, size, material, price, lead time, and a dozen other attributes. Attributes that are missing from the catalog are missing from the comparison.

4. The API exposure

Some agents read public web pages. Some read structured feeds or APIs provided by merchants. As agent programs expand — Perplexity Buy with Pro, ChatGPT agent merchant integrations — the structured-feed path becomes more important. ACO programs consider both.

Why catalog-data readiness, not content, is the ACO primitive

A content-first optimization program writes better PDPs, better category pages, better blog posts. That work helps human readers and, by extension, the engines that cite on behalf of human readers. ACO asks a different question: when an agent reads the structured data behind a PDP, is the data complete enough for the agent to act?

Most Shopify catalogs fail this test at scale. Common failures eCommerce Insights observes in D2C audits as of Q1 2026: variant-level attributes missing on 30 to 60 percent of SKUs, GTINs missing on long-tail products, metafield coverage concentrated on best-sellers, availability flags correct at the parent-product level but inaccurate at the variant level, priceValidUntil set to a default far-future date that engines treat as unreliable.

None of these are content problems. They are structured-data hygiene problems. ACO is largely the discipline of fixing them.

What agentic shopping looks like in Q1 2026

Three observable patterns as of Q1 2026, based on eCommerce Insights's use of these products in manual evaluations.

Pattern 1: Research-to-recommend

A user asks a conversational agent for product recommendations. The agent returns a curated list with citations. The purchase is completed by the human, usually on the merchant's site. This is the most common agent pattern today and is mostly an AEO problem.

Pattern 2: Research-to-cart

The agent compiles a shortlist, then adds items to the merchant's cart via a browser automation or integration. ChatGPT's Operator and comparable products sit here. ACO matters because the agent reads the PDP and structured data to decide what to add.

Pattern 3: Research-to-purchase

The agent completes the transaction. Perplexity Buy with Pro is the most public example as of Q1 2026; availability is limited. ACO is the core optimization discipline here because the agent never loads a human-readable page at checkout.

The agents that matter: Operator, ChatGPT, Perplexity, Rufus, Sparky

Ordered by D2C relevance. The list changes quarterly; verify against current engine disclosures.

1. ChatGPT agents and Operator. Operator, OpenAI's browsing agent, can add to cart and complete purchases on behalf of subscribers. ChatGPT Shopping surfaces products inside the chat flow. eCommerce Insights tracks both as separate surfaces.
2. Perplexity Buy with Pro. A checkout-complete flow for a curated merchant set, early access as of Q1 2026. Significant for D2C brands in the program; less so for non-members.
3. Gemini shopping flows. Integrated with Google's merchant feeds; the agent experience is closer to Google Shopping's existing surface than to Operator.
4. Rufus. Amazon's assistant. Relevant only if the brand also sells on Amazon, where Rufus reads the Amazon catalog rather than the merchant's Shopify store.
5. Sparky. Walmart's assistant. Same scope limitation as Rufus.

A Shopify-first D2C brand should prioritize ChatGPT and Perplexity for ACO work. Rufus and Sparky matter for brands that list on those marketplaces; the optimization work on the Shopify store does not improve Rufus or Sparky results.

A catalog readiness checklist

The seven checks eCommerce Insights runs when scoring a catalog for ACO readiness.

1. Product schema on every PDP. Not a sample. Not the top-100. Every SKU.
2. Variant-level schema where variants exist. A medium and a large of the same shirt are different SKUs for the engine.
3. GTIN, MPN, and SKU populated. Missing identifiers reduce agent confidence in matching products to buyer intent.
4. Offers block complete. price, priceCurrency, availability, priceValidUntil, url. All five.
5. Availability accuracy. The availability value on the PDP matches the real inventory state within a few hours, not a few days.
6. Metafield coverage for spec attributes. Size, material, weight, dimensions, care instructions, country of origin — whatever the category requires.
7. Merchant feeds. If the brand participates in agent merchant programs (Perplexity, Google Merchant Center, ChatGPT's early merchant flows), the feed matches the PDP.

Schema requirements for agentic commerce

The minimum Product JSON-LD block eCommerce Insights recommends for a Shopify PDP aiming at ACO readiness:

Required fields: name brand sku gtin (or mpn) description image offers. Inside offers: price priceCurrency availability priceValidUntil url.

For products with variants, emit one Product record per variant, linked via isVariantOf to the parent product. Schema.org documentation at schema.org/Product covers the full property set; Google's Product structured data guidelines are practical for Merchant Center work.

eCommerce Insights's Product Schema Generator produces a compliant block from a Shopify product record in a single click. For automated push across the catalog, eCommerce Insights's PDP optimization handles the diff-and-approve workflow.

Price and availability freshness

Agents need current prices and real availability more than humans do. A human who sees a slightly stale price will tolerate the surprise at checkout; an agent comparing products against a budget will deprioritize any product whose price or availability it cannot trust.

Three practical rules eCommerce Insights observes:

First, priceValidUntil should match the real promotional cadence, not a default of "one year from today." Engines discount the confidence of a Product block whose priceValidUntil is implausibly far in the future.

Second, availability flags should propagate from Shopify's inventory system to the PDP within minutes, not hours. For brands using third-party inventory tools, verify the propagation delay.

Third, out-of-stock products should emit availability "OutOfStock" rather than a cached "InStock" value. An agent that recommends an unbuyable SKU erodes its own credibility and the merchant's trust score across sessions.

A SKU whose price the agent cannot trust is a SKU the agent does not recommend. Freshness is an ACO signal, not a nice-to-have.

Review-source grounding for agentic shopping

Agents ground their confidence in third-party review data. A product with visible aggregate review counts from Shopify-native apps (Judge.me, Yotpo, Okendo) and cross-referenced coverage on independent review sites (Trustpilot, category publications) is more likely to be recommended than one with no third-party grounding, based on eCommerce Insights's observation of Perplexity citation patterns through Q1 2026.

For ACO specifically, the structured expression of reviews matters. An AggregateRating block with ratingValue, reviewCount, and a sampling of Review sub-records gives the agent machine-readable grounding. A marketing page that claims "5-star reviews across the board" without structured expression is invisible to the agent.

Where ACO will head next

Three directional bets eCommerce Insights is watching through the rest of 2026. None are predictions; all are uncertain.

Merchant-specific feeds expand

More engines will publish structured-feed specs for merchants. Today Google Merchant Center is the best-developed; Perplexity and ChatGPT have early merchant programs that may converge on similar formats. Brands with clean Shopify data will have an easier time adapting.

Agent-to-agent protocols emerge

Standards bodies are beginning to discuss agent-to-agent communication protocols for commerce — protocols that let a buyer's agent query a merchant's agent directly. Any such standard will elevate structured catalog data further.

Review grounding shifts toward verified-purchase signals

As agents complete more purchases, verified-purchase review aggregation becomes a stronger grounding signal than open review counts. Brands investing in Shopify-native verified review tools will benefit.

What to do this quarter

A 90-day ACO plan for a Shopify D2C brand:

1. Week 1. Run the seven-point catalog readiness checklist above. Score each SKU.
2. Weeks 2 to 4. Fix variant-level schema and GTIN coverage across the top-100 revenue SKUs.
3. Weeks 5 to 7. Audit price and availability freshness. Repair propagation gaps between Shopify admin and the PDP.
4. Weeks 8 to 10. Enroll in merchant programs where relevant: Google Merchant Center, Perplexity merchant integration if eligible.
5. Weeks 11 to 13. Baseline citation measurement across ChatGPT Shopping, Perplexity Shopping, and Gemini. Establish weekly monitoring.

Key takeaways

• ACO was coined by and describes catalog readiness for AI shopping agents.
• The ACO primitive is structured data, not content — schema, variants, metafields, availability.
• For Shopify D2C, ChatGPT agents and Perplexity Buy with Pro are the agents that matter first.
• Price and availability freshness are first-class ACO signals; stale values reduce agent confidence.
• A catalog audit using the seven-point readiness checklist is the fastest way to baseline a program.

Related guides

Use the tools