Criteo introduced its Agentic Commerce Recommendation Service on February 5, 2026, aiming to provide AI shopping assistants with commerce-grade product recommendations grounded in real purchasing behavior.
The service was designed to help AI assistants deliver more accurate and transaction-ready results than approaches based solely on product descriptions.
According to Criteo, internal testing showed the service improved recommendation relevancy by up to 60% compared with third-party methods that rely only on catalog data.
The system was built on Criteo’s commerce intelligence, which includes signals from 720 million daily shoppers, roughly $1 trillion in annual transactions, and 4.5 billion product SKUs.
The service is available through Criteo’s Model Context Protocol and connects AI assistants directly to merchant inventory. When a consumer submits a shopping request, the AI assistant queries Criteo’s service, which applies real-world shopping and purchase signals to filter and rank products based on intent, availability, and popularity.
The output is a curated shortlist of recommendations rather than a raw product feed.
Criteo said the service supports both exploratory and product-specific queries and can return complementary items where relevant. Testing with a major large language model platform began in 2025.
Image Credit: Criteo
Why This Matters Today
You are seeing AI shopping assistants move closer to handling discovery and purchase decisions end to end.
As these assistants scale, the quality of product recommendations becomes a limiting factor, especially when models rely only on publicly available descriptions that lack real performance signals.
Criteo’s approach emphasized transaction data over scraped content. By using real shopping and purchase behavior, the company aimed to give AI assistants signals that better reflect what consumers actually buy, not just what products claim to offer.
That distinction matters as AI systems increasingly influence high-intent shopping decisions.
The launch also pointed to a shift in how commerce infrastructure is exposed to AI platforms. Instead of crawling catalogs, AI assistants can query services designed to translate intent into ranked, purchase-ready results.
For retailers and brands, this creates a path to participate in AI-driven shopping without handing over raw data or losing control of inventory and pricing.
As agentic commerce evolves, recommendation infrastructure may become as critical as payment and logistics. Criteo’s service positioned the company as a backend provider for AI shopping experiences rather than just an advertising platform.
Our Key Takeaways:
Criteo launched an agentic commerce recommendation service for AI shopping assistants. The service used real transaction data to improve product relevancy. Criteo reported up to a 60% improvement in relevancy during internal testing.
The launch highlighted growing demand for commerce-grade infrastructure in AI-driven shopping.
- Criteo introduced a recommendation service designed for AI shopping assistants.
- The system uses real shopping and purchase signals, not just product descriptions.
- Internal tests showed up to a 60% increase in recommendation relevancy.
You may also want to check out some of our other tech news updates.
Wanna know what’s trending online every day? Subscribe to Vavoza Insider to access the latest business and marketing insights, news, and trends daily with unmatched speed and conciseness. 🗞️
