Blog

Powering the future of AI commerce: How Mirakl is accelerating the agentic era

Ashley Parker - July 17, 2025
image

A new chapter is unfolding in commerce — one shaped by AI, fueled by innovation and driven by those reimagining how products are bought and sold. At the 2025 Mirakl Summit in New York City, industry leaders convened to explore the next wave of marketplace growth, with AI taking center stage.

In a conversation with Omni Talk Retail, Mirakl Chief Data and AI Officer Anne-Claire Baschet pulled back the curtain on how AI is reshaping the industry and how Mirakl is not only investing in AI innovation, but embedding it where it drives the most value.

Building AI that solves real problems

Baschet emphasized that AI’s value lies in solving everyday challenges, not simply keeping up with trends.

“My role is to help our customers identify where AI creates value and integrate it seamlessly into their workflows, whether for sellers, operators or across 1P and 3P models,” she explained.

That same philosophy is applied internally. 

According to Baschet, 91% of Mirakl employees now use AI in their day-to-day work, and nearly one-third have built their own agents. 

These tools range from simple automations to solutions that streamline complex tasks. This hands-on adoption helps reduce uncertainty around AI and empowers teams to work more efficiently and creatively.

AI that operates behind the scenes

Mirakl’s AI capabilities are helping remove friction from some of the most complex and time-consuming areas of marketplace operations, particularly product catalog onboarding and enrichment.

To tackle these challenges, Mirakl developed Catalog Transformer, an AI-powered tool that automates key tasks such as product categorization, attribute mapping and taxonomy alignment.

By eliminating manual bottlenecks, Catalog Transformer enables faster, more scalable seller onboarding while ensuring product data remains clean, consistent and optimized for discovery.

“Problems that once felt overwhelming are now being resolved quickly with AI,” Baschet noted. “That’s when the real impact becomes clear.”

Embracing the agentic commerce era

Looking ahead, Baschet sees “agentic commerce” as the most transformative trend on the horizon. In this emerging model, AI agents, rather than consumers, will increasingly interact with digital platforms, search for products and make purchasing decisions.

To support this evolution, Mirakl is investing in commerce-specific AI models, fine-tuned to outperform general-purpose tools on retail-centric tasks. This includes improving the speed and accuracy of product data enrichment and enabling more dynamic search experiences.

A critical part of this strategy is the adoption of multimodal capabilities, which combine both image and text data. 

“If a product description says ‘blue dress’ but the shopper is searching for ‘dark blue,’ the AI agent can interpret that through the product image,” she noted. “High-quality visuals are becoming just as important as detailed copy.”

This blend of data sources is improving precision and relevance, especially as search queries grow more context-rich and personalized.

As consumers adopt new ways to express their intent, whether through agents, conversational interfaces or more visual search inputs, the ability to serve relevant, structured and high-quality content will become a key competitive advantage.

Transforming retail media and product discovery

Beyond catalogs, AI is reshaping how retailers approach media and search. With more advanced targeting capabilities, advertising operations are becoming more efficient, allowing retailers to unlock new revenue streams and better integrate third-party sellers into their campaigns.

Simultaneously, the pathways for product discovery are shifting. Search behavior is evolving, and traditional routes to visibility, like brand recognition or paid placement, are no longer guaranteed. Intelligent discovery is now being mediated by new technologies, where product relevance and content quality carry more weight than ever.

“For years, brands have invested heavily in awareness. Now, discovery may happen in environments where visibility isn’t guaranteed,” Baschet said. “But that also opens the door to entirely new strategies for reaching customers.”

Retailers and sellers alike must adapt by ensuring product data is not just complete, but context-aware and optimized for discovery across emerging digital environments.

From innovation to execution

Baschet’s perspective on AI is deeply rooted in execution. She notes that, at Mirakl, innovation isn't just about experimenting with what's possible, but about building scalable tools that address the operational pain points which hold back businesses.

As the industry enters the next era of commerce, success will depend on digital readiness and the ability to integrate intelligent systems that drive measurable value from day one. This shift isn’t theoretical or far off — it’s already reshaping the way businesses operate.

Learn more about agentic commerce and what it means for the future of online buying, here.

image
Ashley Parker,
Corporate Marketing Co-Op

Related content

An infographic titled "Artificial intelligence and product data" presents a pink, double-breasted blazer at its center. Surrounding the blazer are four boxes containing information about the product.

A box in the top left corner is labeled Standardized Color Code and shows a pink swatch with the hex code #B02967.

The top right box, labeled Fun fact, notes that the blazer was "Worn by Rachel in Friends S02E12."

In the bottom left, the Extracted Attributes box lists:

- Age Group: Adult

- Coat Length: Mid Length

- Coat Style: Blazer

- Gender: Women

- Apparel Tax Class: General Clothing

The bottom right box, titled Features & Benefits, highlights:

- Double-breasted design

- Stylish and modern fit

- Perfect for both casual and formal occasions

The box also includes a section for Available translations with the flags of France, Spain, Italy and China.

A final box in the middle left, labeled Suggested styling, advises pairing the blazer with "slim-fit trousers or a pencil skirt for a formal look.

AI, agentic commerce and the need for impeccable product data

A screenshot of an AI agent's response to a user's query. The agent has identified a fuchsia double-breasted jacket, linked it to one worn by Rachel in the TV show Friends and provided details for purchasing the jacket from a store called "Velvet & Vogue." The agent also indicates the product's availability in various sizes and an estimated delivery date.

From physical to agentic: The next great commerce shift

An illustration of an agentic commerce shopping experience. At the top, a user leverages an AI agent with the prompt "I want to buy that fuchsia jacket I saw on Friends." Below the chat bar, the AI agent responds: "This is a fuchsia jacket similar to one Rachel wore in Friends TV show S02E12. It is available in your size at these stores." The rest of the screen shows product listings for fuchsia jackets from different stores, including Urban Outfitters and Kohl's, with prices and available sizes.

What is agentic commerce?