A year ago, ConsoTrust, a specialist in artificial intelligence for compliance, processing and analysis of product information, joined the AGENA3000 group. Meet Julien DOYEN and Julien BOYER, co-founders of ConsoTrust, who are at the heart of product data quality.

What inspired you to found ConsoTrust, and where did the idea come from? 

« ConsoTrust wasn't our first idea. As parents of allergic children, our first step was to create an application to help consumers with food allergies and intolerances to find products compatible with their restrictions.

When we created this application and requested data from manufacturers and distributors, we realised that the quality of the data was so poor that we couldn't launch an application without risk.

Having this expertise, we set up checks on allergen information to verify consistency with ingredient lists. These checks were of interest to the manufacturers and distributors who sent us data, and that's how ConsoTrust was born.

Alongside the Allergobox* application, we have created ConsoTrust, a platform for data validation and compliance.

Initially, our scope extended to allergens, as this was our first field of action. Gradually, in response to requests from manufacturers and distributors, we moved on to other attributes: regulatory information, logistical information, etc. Our field of action has thus become much broader than the allergens we started with.

We have been materialising this monitoring activity through ConsoTrust since 2019. »

*Allergobox : free web portal dedicated to food allergies and intolerances.

Can you tell us about the major issues that ConsoTrust is addressing through its services and who they are aimed at? 

« With ConsoTrust we can list 3 major challenges: 

• Data quality : reliable data that complies with regulations. If we're talking about an allergen, for example, it must not put the consumer at risk.
• Productivity : manufacturers still manage data very manually, and there is a lot of data entry. The aim is to be able to automate processes such as data acquisition by extracting documents, for example, or to automatically calculate enrichments so that operators don't have to enter them.
• Providing services to the company's business lines: for benchmarking, for improving revenue, for monitoring the competitive environments of its products, for stimulating innovation, etc. These are use cases that can only be offered if the data is of good quality. »

In practical terms, what are the consequences of poor product data quality?

« The impact of poor data quality is exponential, as we have to manage more and more data. What's more, today there are more attributes attached to a product than there were ten years ago, so the more attributes there are, the greater the possibility of making an error, and the greater the impact of these errors.

There are various types of impacts :

• The regulatory impact : the information must be accurate, within the framework of the INCO* regulations, with the risk of penalties from the public authorities, and therefore a financial risk
• Health risk/consumer risk : with allergens, the risk can go as far as death, and this is every manufacturer's worst fear
• Image risk : having a product with information that doesn't match is not very serious. If you forget to declare an additive, either by oversight or inadvertence, consumers will think that you're not being transparent.
• Commercial risk : consumers can use this information to make choices. For example, with the Yuka consumer application, if my information is wrong, I may get a lower score. »

*INCO : The INCO Regulation (Food Information for Consumers) is a European Union (EU) regulation that sets out the requirements for food information aimed at consumers. 

Are you seeing a growing interest from companies in the quality of their product data?

« Yes, there is a growing interest, but above all it's an obligation, because it can be penalised by the public authorities and by customers (distributors). If the data is incorrect, penalties can be applied. »

How do you explain the recurring presence of errors in product sheets?

« We can explain this by the data entries. The quality of product data depends on the reliability of its source. 

The data passes through different systems until it ends up on an e-commerce site. As a result, data can deteriorate, become out of sync and be altered by manual intervention. 

Two factors can make data erroneous. On the one hand, data entry (data entry at source, out-of-date data, erroneous entries, attribute inversions) requires more and more technical knowledge: the more information there is to enter, the more errors can be made. 

On the other hand, the more systems the data passes through, the more likely it is to be wrong at the end, with the risk of desynchronization and missed updates. » 

Finally, from the point of view of liability between manufacturers and distributors, who is responsible in the event of allergen errors (for example) on a Drive distributor site?

« It is the owner of the data, i.e. the manufacturer, who is legally responsible, but all parties are concerned, because distributors and suppliers are linked to each other. 

Errors can also come from the distributor, because when the data reaches the distributor, it passes through different systems and there can be desynchronization. It is very common for the manufacturer to send the right data at the outset, but it is not the same data that ends up on the distributor's drive. For example, in the case of product sheet versioning, the retailer may not consider the latest version or update the associated visuals. 

With ConsoTrust, we have a lot of value to bring, offering all-in-one solutions that replace these different systems. » 

Artificial intelligence is on everyone's lips. What sets you apart from your competitors? 

« We're very advanced in the field of artificial intelligence, as we've been doing it almost since the start of the venture. We have our own models at the cutting edge of its application in the world of mass retailing. At ConsoTrust, we talk less about AI than those who do less... In fact, we talk more about the business and the challenges facing our customers, because AI is only a means to an end - what matters to the customer is the result. 

For example, on the extraction side, we have high-performance models that are reliable in real time. So, our priority is to respond faster and faster to more and more customer challenges thanks to AI. » 

What are ConsoTrust's priorities for development in 2026? 

Since September 2025, the ConsoTrust teams have been fully integrated into the AGENA3000 teams. Their expertise strengthens the group's Product Ecosystem Automation area, which includes PIM, PDM and AI solutions.