Criteo is betting that the next battleground in digital commerce won’t be search bars or product listing ads—it’ll be AI shopping assistants.
The commerce media company (NASDAQ: CRTO) has introduced its new Agentic Commerce Recommendation Service, a system designed to feed large language model (LLM) platforms and retailer chatbots with transaction-ready product recommendations powered by real shopping data—not just scraped product descriptions.
As AI assistants increasingly guide how consumers discover and compare products, Criteo’s pitch is simple: generic product feeds won’t cut it. Commerce‑grade intelligence will.
Why This Launch Matters
LLM platforms are rapidly evolving into conversational shopping guides. Retailers, meanwhile, are building AI chatbots into their own digital storefronts. But most AI systems rely heavily on publicly available product metadata—titles, descriptions, specs—without access to real‑time shopping behavior signals.
That creates a gap between relevance and reality.
Criteo’s new service attempts to close that gap by plugging AI assistants directly into its commerce intelligence layer, which is built on:
- 720 million daily shoppers
- $1 trillion in annual transaction data
- 4.5 billion product SKUs
According to Criteo’s internal testing, recommendations powered by its system delivered up to 60% higher relevancy compared to third‑party approaches based solely on product descriptions.
In a world where AI assistants increasingly influence purchase decisions, that kind of lift could materially impact conversion rates—and ad dollars.
How It Works
The Agentic Commerce Recommendation Service operates through Criteo’s Model Context Protocol (MCP), acting as a bridge between AI shopping interfaces and live merchant inventory.
Here’s the flow:
- A consumer asks an AI assistant for a product recommendation.
- The assistant queries Criteo’s service.
- Criteo applies real‑world commerce signals—purchase data, popularity, availability, and intent modeling.
- The system returns a curated shortlist of transaction‑ready products.
- The AI assistant presents results and can support add‑to‑cart or checkout within the experience.
Instead of dumping raw catalog data into an LLM, Criteo filters and ranks products based on observed shopping behavior. The result is meant to reflect what people actually buy—not just what product copy suggests.
The system also handles both exploratory queries (“best running shoes for beginners”) and product‑specific searches, while recommending complementary items when relevant.
Commerce Intelligence as Competitive Moat
Criteo has long positioned its commerce data graph as a differentiator. This launch reframes that asset for the AI era.
“The real competitive advantage in agentic commerce will come from access to high‑quality commerce data at scale,” said CEO Michael Komasinski.
The broader implication: as AI assistants reshape the discovery funnel, the power may shift toward platforms that control transactional intelligence rather than just ad placements.
Retail media networks, marketplace giants, and walled gardens already sit on valuable first‑party commerce data. Criteo’s move suggests independent commerce media players want to insert themselves directly into AI‑driven purchase workflows—before those ecosystems become closed loops.
The Rise of Agentic Commerce
“Agentic commerce” refers to AI systems that don’t just recommend products but can take actions—comparing options, adding items to carts, even completing purchases autonomously.
OpenAI, Google, Amazon, and other players are exploring AI‑driven shopping experiences. If conversational commerce becomes mainstream, recommendation infrastructure will become mission‑critical.
That raises key questions:
- Who supplies the ranking logic?
- Who controls the data layer?
- Who captures the monetization opportunity?
Criteo’s service positions the company as an infrastructure provider rather than just an ad network—embedding itself within AI‑powered commerce flows.
Early Testing and Expansion Plans
Criteo says it began testing the service with a major LLM platform in 2025 and is expanding trials to additional LLM providers, retailers, and brands.
The timing is strategic. AI‑powered search and shopping interfaces are still in early development, which means integration partnerships formed now could shape long‑term monetization models.
If successful, the Agentic Commerce Recommendation Service could allow Criteo to participate directly in AI‑native purchase journeys—rather than relying solely on traditional display and retail media ad formats.
The Bigger Picture
As AI assistants move from novelty to mainstream utility, recommendation quality becomes the currency of trust.
Consumers will quickly abandon AI shopping tools that surface irrelevant, out‑of‑stock, or poorly ranked products.
Platforms that can blend conversational fluency with real commerce intelligence will have an advantage.
Criteo’s latest move signals a strategic pivot toward becoming the recommendation engine behind the AI shopping layer. Whether it can secure durable partnerships in an increasingly competitive AI ecosystem remains to be seen.
But one thing is clear: in the emerging world of agentic commerce, data depth may matter more than conversational charm.
Get in touch with our Adtech experts
