Confidentiality, close to the way you work

Lexanon is published by GlobeNews SAS in Rennes, France. We design local pseudonymisation tools for professionals working with confidential documents, bringing AI closer to their work and giving them control over the information they share.

Our experience in legal technology and the needs of lawyers and legal professionals guides this approach. Data protection, consistency across documents and review before sharing are requirements we want to make accessible in every organisation's daily work.

Protecting data takes time and attention, two resources already in high demand at work. We believe tools should shoulder their share of that effort. Protection that people can adopt for the long term must respect teams' habits, constraints and pace. This is how we see pseudonymisation: a digital hygiene practice that fits naturally into everyday work. Every interruption and extra step informs our design choices.

Making data protection understandable

Ease of use must go hand in hand with an understandable process. Everyone should be able to see what is protected, what is shared and what still needs checking. We want people to retain their judgement and control over their exchanges. Trust grows when technical choices and their limits are explained clearly enough for everyone to make an informed decision.

Local processing also reflects our view of autonomy. Working on your own computer reduces the need to entrust information to others and rely on remote infrastructure. We aim to use servers only where necessary. This leads us to consider what a machine can achieve on its own and to use its resources carefully.

This autonomy should extend beyond organisations with IT departments and powerful computers. A freelancer or small business should be able to protect their information with the computer they already own. We therefore devote part of our work to reducing our models' memory and computing requirements. Making protection accessible also means considering hardware costs and avoiding the need to replace equipment to adopt new tools.

Tools that prioritise autonomy

Our commitment to open source follows the same pursuit of autonomy. Being able to study software, adapt it and discuss its design with other developers helps build tools we understand better. We want to contribute to this exchange of knowledge and support an ecosystem in which more people can develop, understand and use AI on their own machines.