Adobe adds AI discovery tools to Commerce catalogue

Adobe has added tools to Adobe Commerce to improve product discovery on large language model platforms. The features are available natively across Adobe Commerce deployment models.

The update centres on what Adobe calls Catalog Agent, which adds structured product information from a merchant's existing catalogue to a machine-readable layer for AI crawlers and LLM-driven shopping and search services. It does not alter the customer-facing storefront, product imagery or buying journey.

That structured layer can include product names, attributes, specifications, compatibility, availability, pricing and related catalogue data. The aim is to help AI systems interpret products more accurately when consumers ask conversational shopping questions through services such as ChatGPT, Gemini, Claude and Microsoft Copilot.

The move reflects a broader shift in online retail traffic as consumers use generative AI tools earlier in the buying process. According to Adobe Digital Insights, traffic from AI sources to US retail sites rose 125% between April and June compared with the same period a year earlier, following a 693% year-on-year increase during the November-to-December 2025 holiday shopping period.

Catalog changes

The new functions enrich product names, descriptions and use-case phrases directly within the Adobe Commerce product catalogue. By making those changes at the source, Adobe aims to keep product messaging consistent across storefronts, advertising channels, marketplaces and AI-driven discovery services.

For merchants, that means a single catalogue can feed multiple channels with the same underlying product story, rather than relying on different descriptions and formats for each destination. In turn, AI systems get more structured context when generating recommendations, comparisons and answers to product questions.

The tools are also intended to reduce the need for custom integrations to expose catalogue data to AI applications. Instead, Adobe Commerce uses existing commerce services, including inventory, pricing, attributes and product relationships, so AI applications can retrieve current information from the platform.

Discovery shift

The launch comes as retailers and commerce software providers adapt to a search market no longer defined solely by web search engines, on-site search bars and marketplace listings. Conversational AI interfaces are becoming another route for consumers to research products before visiting a retailer's own site.

Adobe argues that these interfaces change the basis of discoverability. Rather than depending mainly on keywords, merchants also need catalogue data that AI systems can parse and reason over when a shopper asks for a recommendation, comparison or compatibility check.

Examples cited by Adobe include prompts asking for "lightweight trail running shoes suitable for marathon training", comparing laptops for video editing or finding accessories compatible with a camera. In those cases, the quality and structure of product data can affect whether a product appears in the response.

Native feature

The LLM discovery functions are built into Adobe Commerce rather than offered as a separate platform or external connector. That may appeal to existing users who want to test AI-driven product discovery without replacing their current commerce system or duplicating product data in another service.

Adobe framed the update as part of a wider push into what it describes as agentic AI for commerce. In practice, the immediate change is a new way to expose catalogue information so AI systems can access richer, governed data from merchants already using Adobe Commerce.

For business users, the commercial rationale is straightforward. If more shopping journeys begin inside AI assistants, retailers may need to ensure their products can be understood and surfaced there, not just ranked in conventional search results or promoted through marketplaces.

Technical teams, meanwhile, face a separate challenge around data quality and consistency. Product information often sits across multiple systems and channels, and any mismatch in specifications, names or availability can make it harder for automated systems to present accurate answers.

Adobe's approach is to keep the product catalogue as the main source of truth and expose that data in structured form for AI consumption. That leaves the storefront experience unchanged while adapting the underlying product data for a retail environment where machine interpretation plays a larger role in discovery.