On April 27, Commerce (Nasdaq: CMRC), parent of BigCommerce and Feedonomics, announced that merchants are syndicating product catalogs to agentic discovery channels through Feedonomics Agentic Catalog Exports (ACE), a new enterprise service. ACE delivers optimized catalogs to OpenAI/ChatGPT, Google AI surfaces including Gemini, Microsoft Copilot, PayPal, Stripe, Perplexity, and Amazon. Dell is the named launch customer, with approximately 7,000 SKUs (laptops, desktops, servers, monitors, accessories) prepared for AI-driven discovery. The service is enterprise-only, with self-service expansion planned.
Most coverage framed this as retail news. SalesTechStar, GlobeNewswire, and Stock Titan led with Dell as a brand-discovery story. For B2B teams running product catalogs that include configurable SKUs, gated pricing, and contract-tied availability, the actual news is different: AI-driven discovery is a feed-management discipline now, and the schema retail uses doesn’t map cleanly onto a B2B catalog.
Our read: ACE confirms what G2 Answer Economy data and Trustpilot’s ChatGPT citation share already indicated. AI search and AI-driven product discovery is being operationalized at the feed layer, not the SEO layer. The launch is a useful trigger for B2B teams to audit how product data shows up in agentic surfaces and where schema gaps will produce the worst answers when an AI agent compares your offering to a competitor’s.
Key Takeaways
- Feedonomics Agentic Catalog Exports (ACE) launched April 27, 2026 from Commerce (Nasdaq: CMRC), parent of BigCommerce and Feedonomics.
- ACE syndicates product catalogs to OpenAI/ChatGPT, Google Gemini, Microsoft Copilot, PayPal, Stripe, Perplexity, and Amazon agentic surfaces.
- Dell is the named launch customer with ~7,000 SKUs structured for ChatGPT product representation; the service is currently enterprise-only.
- B2B catalog data has structural gaps consumer schemas don’t cover: configurable SKUs, gated pricing, contract-tied availability, channel-specific terms.
- The right operator move isn’t immediate ACE adoption — it’s auditing how existing feeds render in agentic surfaces and identifying where B2B-specific schema gaps will produce wrong answers.
What Feedonomics Actually Shipped
ACE is an enterprise service that takes a merchant’s product catalog and prepares it for syndication to “agentic discovery channels,” Feedonomics’ phrase for AI-powered product surfaces. Currently supported destinations include OpenAI/ChatGPT, Google AI surfaces (including Gemini), Microsoft Copilot, PayPal, Stripe, Perplexity, and Amazon. The press release describes ACE as building on Feedonomics’ existing data transformation, enrichment, and syndication capabilities, extending them into agentic protocol-driven destinations.
The launch customer matters more than the service description. Dell, through Feedonomics, has prepared approximately 7,000 product items (laptops, desktops, servers, monitors, accessories) for AI-driven discovery. Paul Mansour, Dell’s global marketing director, framed the move as ensuring Dell products are more accurately and completely represented within ChatGPT
. That’s the B2B-relevant version of what most coverage missed: ACE is not only a retail discovery play. Dell is an enterprise hardware vendor selling configurable products through complex channels.
The roadmap mentions self-service and mid-market expansion. Initially, ACE is enterprise-only. Mid-market and self-service brands will get access later through the existing Feedonomics platform.
Why B2B Catalogs Are a Different Problem from Retail
B2B product data carries structural complexity consumer schemas don’t. A B2B catalog routinely includes: configurable SKUs (a server with 47 component options), gated pricing (no public list price; quote required), contract-tied availability, volume-tiered pricing, region-locked SKUs, and product bundles tied to service contracts (the SKU exists but isn’t standalone-sellable).
Retail catalog schemas largely don’t carry these signals. When an AI agent answers “what’s the cheapest server with 64 GB RAM and 24-core CPU for a 200-person team,” the answer depends on whether catalog data carries the configurability, tier-based pricing, channel-partner restrictions, and bundle dependencies. If the underlying schema is consumer-style (single SKU, single price, single availability), the AI agent’s answer will be wrong in ways that close deals nobody can fulfill at the price the AI quoted.
The parallel to Microsoft’s Universal Commerce Protocol launch in B2B is direct. Microsoft positioned UCP for retail but mentioned applicability to “any productized offering.” ACE is the same shape: the syndication infrastructure is real, but B2B-specific schema work is the operator’s responsibility, not the platform’s. AWS’s May 7 AgentCore Payments preview closes the third layer of the same agentic-commerce stack: UCP standardizes the data, ACE syndicates it to AI surfaces, and AgentCore Payments lets the agent transact against it without a human in the loop. Shopware’s connected agentic-commerce architecture puts the context, workflow, storefront, and payment layers around that feed-level foundation.
What B2B Operators Should Do This Quarter
Audit how your catalog renders in current AI surfaces. Search your top 5 product names in ChatGPT, Gemini, Copilot, and Perplexity. Check what each agent says about pricing, configuration, availability, and channel access. Document where the answers are factually wrong, reference outdated SKUs, or imply availability that doesn’t exist. The audit takes a few hours and produces the priority list for whichever feed-management approach you adopt. The audit should now include post-purchase trust, because 31% say an autonomous purchase can reduce their likelihood of returning even when it succeeds.
Map your B2B-specific schema gaps. Once you’ve audited the rendering, identify which catalog fields are missing or misrepresented: configurability, gated-pricing flags, channel-partner restrictions, volume tiers, region locks, bundle dependencies. These are the schema points consumer-focused syndication tools may not handle natively. The gap list is what you negotiate against, whether the answer is ACE, Mirakl Agentic Activation, ChannelSight, or a custom pipeline.
Treat AI-surface presence as a line item, not a project. The pattern in agentic marketing across 2026 is that AI visibility budgets are getting compressed into existing martech budgets, not expanded. Teams winning this year treat AI catalog readiness like SEO hygiene in 2015: recurring effort, owned by an existing team, measured against citations and accuracy rather than launch projects.
The broader pattern: Mondelez’s agentic commerce strategy, Adobe’s Marketo MCP launch, and Feedonomics ACE all point at the same operating reality. Product and customer data are getting machine-readable across the stack. B2B teams that audit, fix the schema gaps, and treat agentic surfaces as a recurring discipline will compound advantage; teams waiting for a clean vendor package will compound the answers competitors give customers.
Frequently Asked Questions
An enterprise service launched April 27, 2026 by Feedonomics, the data feed optimization arm of Commerce (Nasdaq: CMRC), parent of BigCommerce. ACE syndicates merchant product catalogs to agentic discovery surfaces including OpenAI/ChatGPT, Google Gemini, Microsoft Copilot, PayPal, Stripe, Perplexity, and Amazon. Currently enterprise-only, with self-service expansion planned.
ACE technically supports any productized catalog, and Dell (a primarily B2B enterprise hardware vendor) is the named launch customer with approximately 7,000 SKUs structured for ChatGPT representation. However, B2B-specific schema challenges (configurable SKUs, gated pricing, contract-tied availability, volume tiers, channel-partner restrictions) require additional work beyond what consumer-focused syndication tools handle.
Not as the first move. The first move is auditing how your current product catalog renders in ChatGPT, Gemini, Copilot, and Perplexity to identify where existing answers are wrong, outdated, or reference unavailable configurations. That audit produces the schema-gap list, which is the input you’d need to evaluate ACE, Mirakl Agentic Activation, ChannelSight, or any other agentic-feed approach. Adopting before auditing means you’re paying to syndicate the same gaps faster.
Microsoft’s Universal Commerce Protocol (UCP) is the underlying schema standard for AI-agent-readable product data, co-developed with Shopify and supported in Microsoft Merchant Center. ACE is a managed service that prepares catalogs for syndication to multiple agentic surfaces, both UCP-using and not. They operate at different layers: UCP is the data format; ACE is the pipeline that gets a catalog into that format and others, then delivers it to AI agents at scale.
