AI: Threat or Opportunity? What a Translator Learned from Kodak’s Downfall

AI: Threat or Opportunity? What a Translator Learned from Kodak’s Downfall

As AI began to reshape the translation industry, Paula Reichenberg decided to explore how this technology could reinforce rather than replace specialised expertise. With research partners and support from Innosuisse, this idea evolved into a series of innovation projects, a new company, and new expertise anchored in Fribourg.

Paula Reichenberg founded Hieronymus – Translations by Lawyers for Lawyers in 2007. The company specialises in legal and financial translations, combining linguistic expertise with an in-depth understanding of complex legal matters.

Over time, however, advances in machine translation began to change the market. Clients increasingly started handling standard translations themselves, challenging the company’s established business model.

Rather than viewing AI as a threat, Paula started exploring how it could complement the expertise developed within Hieronymus and open up new opportunities. For her, this meant recognising early on that technological change had to be embraced rather than ignored.

“During my MBA, I kept one little phrase in mind: ‘Click-Clack Kodak.’ It was shorthand for how quickly a market leader can lose relevance if it fails to embrace and harness a new technology. When machine translation started changing our market, I did not want Hieronymus to make that mistake”, explains Paula.

When businesses and researchers join forces

To explore this question, Paula needed expertise that was not yet available within the company. Robert van Kommer, an ICT innovation mentor at Innosuisse, introduced her to Professor Jean Hennebert and the iCoSys Institute at the School of Engineering and Architecture of Fribourg (HEIA-FR).

The partners brought different but complementary strengths to the table. Hieronymus understood legal and financial language, the needs of its clients and the realities of handling confidential information. The research team contributed its expertise in natural language processing, machine learning and data engineering.

An idea with potential?

Their first Innosuisse-supported innovation project began in 2020. Known as Datalambic, it explored how specialised linguistic data could be prepared and used to develop a secure neural translation solution for the legal and financial sectors.

The aim was to automate and accelerate parts of the translation process while keeping professionals in control where legal nuance required irreplaceable human expertise. The collaboration laid the technological foundations for LexMachina, a specialised translation solution adapted to the needs of Swiss legal and financial experts.

Innosuisse support gave the partners the opportunity to test an ambitious idea whose technological feasibility and market potential still had to be demonstrated. It allowed Hieronymus to work with a scientific partner and acquire new capabilities without having to build all the required expertise alone.

Moving forward with support from Innosuisse

The first project was followed by a second collaborative innovation project, again supported by Innosuisse. In 2022, Neur.on AI Solutions SA was founded in Fribourg as a spin-off of Hieronymus to continue developing and commercialising the technology.

Neur.on subsequently developed Corrext, a platform designed to support legal and financial professionals throughout the translation and revision process. Rather than removing human involvement, the solution helps users identify passages requiring particular attention and combine automated processing with specialised legal and linguistic expertise.

In 2024, a third stage began with a Start-up Innovation Project supported by Innosuisse. The different projects accompanied the development from an initial research question to a technology company preparing its solution for the market. According to Alliance Innovation, which supports companies in developing research collaborations and Innosuisse projects, the objectives established through the Innosuisse proposals continued to serve as a roadmap for the company.

An interdisciplinary team as the key to success

As technology developed, Neur.on needed people who could bridge several fields. By October 2024, the company had grown to a team of 11, bringing together expertise in linguistics, law, engineering and data science.

Several students who had worked on the innovation projects were later recruited by the company. Their transition from applied research into company roles provides a concrete example of how collaboration between research and industry can support the development of highly qualified specialists. For Paula, however, building the right team was not only about bringing different areas of expertise together, but also about creating the conditions for genuine interdisciplinary collaboration.

“Legal AI needs an unusual mix of people: lawyers, linguists, engineers and data scientists who are outstanding in their own field, modest about what they do not know, and sincerely open to learning from each other. Once that human neural network starts functioning, you do everything you can to keep it together. That was one reason we decided to raise capital: once you have the right team, you know you can build the product”, she emphasises.

Swiss roots that endure and make a difference in the market

In 2026, Neur.on entered a new phase of development. The company was acquired by Legal 230, Europe’s leading pure player in legal translation, while its eight-member Legal NLP team joined the European Legal AI scale-up Noxtua.Neur.on has kept its headquarters in Fribourg, and all servers and data flows remain in Switzerland.

Swiss data sovereignty has always been one of Neur.on’s USPs and a core focus of its infrastructure. For Swiss public authorities, law firms handling highly confidential matters, and industrial companies protecting sensitive know-how and trade secrets, keeping data and processing in Switzerland is a major requirement. And it has also proved to be a strong selling point for international legal and financial clients, who value Swiss infrastructure for the confidentiality, legal certainty and control it provides over where their data is stored and processed.

Both Neur.on and Noxtua have maintained their presence in Fribourg, allowing the team to contribute to a wider European platform while remaining based in the region.

“The reason Noxtua came to Fribourg was simple: On ne change pas une équipe qui gagne. Noxtua understood the value of keeping together high performers who had already learned how to work with each other across very different disciplines and wanted to build on that”, Paula points out.