EN / ES Tell us your next step
Back to blog
By Nick 4 min read

Catalog Agent: how to get ChatGPT and Gemini to recommend your products

Adobe Commerce now includes Catalog Agent, an AI agent that enriches your catalog so language models can understand and recommend it. What it does, why it matters and how to get ready.

Product cards connected to an AI agent that sends structured data to a chat with product recommendations

For years, the key ecommerce question was "where do I rank on Google?". Today there is a new, increasingly urgent one: when someone asks ChatGPT, Gemini or Perplexity for a recommendation, does the AI mention your products?

Adobe's numbers show this isn't a question for the future. According to Adobe Digital Insights, traffic from AI sources to US retail sites grew 125% year over year between April and June 2026, and during the 2025 holiday season the jump was 693%.

To respond to that shift, in August 2026 Adobe launched Catalog Agent, a native Adobe Commerce capability designed to make your catalog understandable —and recommendable— by language models.

The problem: LLMs don't "read" your store like a shopper does

A human shopper sees photos, reviews and good design. An AI model needs structured context: what the product is, what it's for, what it's compatible with, what it costs and whether it's in stock.

When that information is thin or inconsistent, two things happen:

  • The AI isn't confident enough to recommend the product.
  • Or worse: it recommends it using made-up or outdated data.

Conversational search makes this even clearer. Nobody types "trail shoes" into an assistant; they write "show me lightweight trail running shoes suitable for marathon training", "compare these two laptops for video editing" or "find accessories compatible with this camera". To answer well, the AI needs far more than a title and a SKU.

Side-by-side of the product page a shopper sees in the store and the structured data an AI reads: name, use case, weight, drop, compatibility, price and availability

What Catalog Agent does

Catalog Agent is an AI agent that enriches product data inside the Adobe Commerce catalog itself and exposes it through a machine-readable layer built for AI crawlers and LLM systems. Specifically:

  • It enriches key attributes: names, descriptions, specifications, use cases, compatibility, pricing and availability.
  • It leaves the shopper experience untouched: the storefront humans see stays the same. The enrichment lives in a layer meant for AI.
  • It works with live data: it builds on services Commerce already has —catalog, attributes, inventory, pricing and product relationships— so AI gets current information instead of inferred or stale data.
  • It's source-first: because enrichment happens at the source, it benefits every channel —storefront, advertising pipelines and marketplaces— and keeps them aligned as products change.

One important detail: it's not an add-on and requires no third-party tools. It's natively available in Adobe Commerce on Cloud (PaaS).

Why it matters for the business

Adobe puts it this way: success no longer depends only on search rankings, but on whether AI systems can understand, reason about and confidently recommend your products.

That translates into concrete benefits:

  1. More visibility in conversational AI experiences.
  2. Better recommendations, because the AI has richer context.
  3. Faster purchase decisions for customers.
  4. Less reliance on keyword optimization.
  5. A solid foundation for what's next: agentic commerce, where AI agents don't just recommend, they buy.

The underlying shift is one of mindset: the catalog stops being "a product database" and becomes a strategic asset.

How to get ready (even if you don't use Catalog Agent yet)

Technology helps, but the quality of the outcome depends on the quality of the foundation. A few steps we recommend:

  • Audit your attributes. Do your products have complete, consistent specifications, or do they rely on free-text descriptions?
  • Think in questions, not keywords. Review how customers would describe what they need and whether your catalog can answer.
  • Clean up compatibility and product relationships (accessories, spare parts, bundles). It's among the things AI values most when recommending.
  • Keep price and stock in sync across every channel.
  • Measure traffic coming from AI to understand how much it already weighs in your business.

In short

AI is already a front door to ecommerce, and it's growing fast. Catalog Agent is Adobe's answer to get your catalog ready for it: more context for AI, the same experience for your customers.

At Improntus we work with Adobe Commerce every day. If you want to assess how ready your catalog is for AI-driven discovery, let's talk.

Source: Adobe Commerce for AI-Ready Product Discovery, Adobe Blog, August 2026.

See all articles Tell us your next step