Manutan automates the management of its product references with AI

Manutan, the European leader in B2B e-commerce specialising in the distribution of equipment and supplies to enterprises and local authorities, takes a significant step in optimising its internal processes with artificial intelligence (AI), automating the creation of its product sheets and thus increasing efficiency and productivity.

To efficiently manage over 850,000 product references and descriptions provided by 4,000 suppliers, Manutan has risen to the challenge of automating a process that was previously manual, time-consuming, and costly, with support from AVISIA, a consulting firm specialising in data, AI, and digital, operating in France since 2007.

This task, which involved compiling, correcting, and harmonising large volumes of data, required significant resources and extensive communication between internal teams and suppliers.

Manutan has developed an innovative AI-based solution using a Large Language Model (LLM). This tool, built on OpenAI integrated with our Azure solution, combined with a training dataset hosted on Snowflake, automatically generates descriptions in line with Manutan’s standards and optimised for search engine optimisation (SEO).

The automation process consists of three key steps:

• Automated file processing: The tool analyses the suppliers' files and generates a specific prompt for each product.
• Attribute generation: By using examples from the database, the tool refines the prompt to enable the AI to produce accurate and consistent descriptions that meet Manutan’s standards.
• Data storage and integration: The generated attributes are stored in the cloud, making it easier to verify and integrate them into the company’s systems.

"The Manutan teams quickly recognised the tremendous potential of Generative AI and, through this project, successfully applied it to a concrete use case that creates significant business value. We are very pleased to have supported them and to be a partner in their ambitions around Data & AI" said Pascal BIZZARI, Managing Director, AVISIA.

The results achieved are significant:

• Considerable time savings: Thanks to this automation, a four-day turnaround per supplier file—representing 88% of the average time spent—has been saved, thus reducing email exchanges and manual adjustments.
• Increased employee satisfaction: The teams, relieved of tedious tasks, can focus on higher-value activities, such as developing relationships with suppliers, improving content quality, or managing new projects.
• Cost reduction: The automated rewriting of descriptions has led to a reduction in the use of external service providers, thereby lowering associated costs by 48%.

"In the B2B e-commerce sector, product descriptions are a critical issue that requires significant time and resources to meet both our customers' expectations and compliance requirements. Being able to rely on AI to automate part of these tasks allows teams to focus higher up the value chain and strengthens the attention they give to the final quality of the product sheet," says Guillaume Duval, Chief Data Officer of Manutan Group.

Through this project, which demonstrates the positive impact of AI on business transformation, Manutan reaffirms its ability to innovate and optimise its operations, further enhancing its competitiveness in the B2B e-commerce market.

About Avisia 
Founded in 2007 with the aim of putting people back at the heart of consulting, AVISIA has become a key player in the field of data and AI over the years.  
With a team of passionate experts, this French consultancy specialises in consulting, integration and implementation of projects combining Data, Artificial Intelligence and Generative AI. Today, it carries out a wide range of assignments in cutting-edge businesses and technologies for leading companies in all sectors.