4 August 2026

AI for the Product Experience

AI has advanced so much that it now supports the product experience and, by extension, product data. This data is most often stored, enriched, and contextualized within a repository—such as a PIM —before being displayed across various distribution channels (stores, e-commerce sites, apps) or outreach channels (print catalogs, posters, etc.).

AI—whether through third-party solutions, custom development, or features directly integrated into the PIM —will play a key role in structuring data. AI will therefore be able to enrich, classify, translate, or even correct product data.

Some concrete examples of AI applications

At“Foussier,” a national distributor of the leading brands in building hardware, power tools, and PPE for the finishing stages of construction, a ChatGPT-based tool was used to organize data extracted from supplier files that contained little information and then enrich it from a marketing perspective using manufacturer product data scraped from the web.

The expected benefit was both an improvement in the quality of the generated data and real-time monitoring of the imported data.

Another use case shared at the“Cdiscount” HubForum to improve the recategorization of product listings. They had already made significant progress on this topic in 2022 with their team of data scientists, who had developed algorithms based on a large volume of data: the number of visitors (over 16 million) and the number of products (over 80 million). The arrival of generative AI enabled Cdiscount and its data teams to accelerate the recategorization process, with 14 million of the 35 million product listings fed into the system. The business impact was immediately felt, with a 30% increase in conversion rates for products that underwent this recategorization, thanks to higher-quality information.
A recent example involves a proof of concept (POC) using a digital asset management (DAM) system to identify and sort images from ZIP files and organize them into folders and subfolders: 3,300 images were sorted with a success rate between 92% and 95%. Google’s AI tools, such as Google Lab, Keras, and TensorFlow, were used to carry out this POC using a database of fashion images. This involved setting up a configuration file with the desired keywords for the search function and creating directories and subdirectories to organize the data.

The findings

Among DATASOLUTION’s PIM/MDM clients, we see AI being used in two phases:

Data structuring and enrichment (including translation): use of AI with a variety of tools, such as those offered by Akeneo (quality optimization, spell-checking, harmonization, etc.)
Generation of content (short/long descriptions, images, tags, etc.) based on this structured data

In both of these cases, it is necessary—from an ethical standpoint and to maintain the brand’s image—to always include a final manual review before publication (and sometimes even between the two stages, with a review of “technical” content prior to the pre-generation stage). These roles are performed by data stewards or data managers prior to publication.