What drove the decision to develop an AI-based use case?

Many companies cannot make use of their business data because it contains personal information. This leads to a stagnation of innovation. We wanted to create a solution that enables data to be used quickly, easily, and consistently across systems and file formats, in a legally compliant way. To accelerate this process, we rely on AI that automatically detects personal data in databases and documents. This allows companies to finally use their data without facing privacy risks – regardless of the system vendor.

 

What were your expectations or requirements for the AI application?

It should reliably detect personal data – even in cases of typos or inconsistent spelling. It must understand the data structure to create a consistent anonymization concept – including the identification of duplicates and relationships across databases and applications to preserve usability. The solution must remain transparent, run within the customer’s network, and truly ensure data protection.

 

What frameworks/methods did you rely on for development?

We built on open-source language models and integrated them into our solution in a way that ensures fast deployment and immediate value for the customer. Integration and fine-tuning were carried out by our in-house development team. We built RAG systems to enrich the knowledge and implemented federated learning concepts to keep learning from client-side deployments into our central AI model.

 

The use of AI solutions must also be accompanied by a change in competences. How have you dealt with this challenge in respect of your employees?

Our employees receive targeted training in the use of AI and apply various AI tools in their daily work. A conscious approach to new technologies is part of our company culture. Openness to change and active change management are essential. Our team sees the added value – whether in increased efficiency in routine tasks or in the development of new business models.

 

In what timeframe did they implement the AI solution in your company/your client‘s company?

The first version was developed in three months. Since then, we have continuously improved it in co-innovation projects with customers.

 

A conclusion to your AI solution:

Our AI helps companies unlock the potential of their business data without violating privacy regulations. It significantly accelerates digital transformation projects by increasing data quality, for testing, training and many other secondary usages during digital transformation projects.

Particularly valuable is that we not only enable new AI projects but also make existing datasets usable in a GDPR- and AI Act-compliant way for existing secondary data usage requirements within organization.

 

One final piece of advice/tip to other entrepreneurs looking to apply AI:

Start small but with a clear use case. Think about data protection from the beginning to avoid pitfalls, when you take AI to productive usage.

 

 

Further information:

https://www.mayadataprivacy.com/appsafe

https://www.mayadataprivacy.com/filesafe

www.mayadataprivacy.de

 

Picture: DIZ