top of page

Harnessing AI to help legal professionals

Founded by four ETH Machine Learning PhDs, DeepJudge sets new standards in context-based understanding and processing of legal documents.

DeepPsy co-founders
Dr. Kevin Roth | Co-founder & CDS, Dr. Florian Schmidt | Co-founder & Head of Research, Dr. Paulina Grnarova | Co-founder & CEO, Dr. Yannic Kilcher | Co-founder & CTO

Could you explain what your startup does in a few sentences?

We at DeepJudge are on a mission to free legal professionals from poring over documents and enable them to focus on the truly strategic aspects of their work.

Founded by four ETH Machine Learning PhDs, DeepJudge sets new standards in context-based understanding and processing of legal documents.

At the heart of our technology is a multilingual AI that can discern the content of structured and unstructured legal text and thus quickly point to important information, surface relevant insights and proof check for inconsistencies. DeepJudge’s AI is used as an out-of-the-box assistant to legal professionals that seamlessly integrates with any document management system, enabling users to review and search through large amounts of documents more accurately while saving significant time and resources.

Your startup works with the power of AI. How do you think this technology will develop in the future? Artificial intelligence (AI) is already transforming businesses and it will continue to contribute to economic growth via boosts in productivity. Just as DeepJudge’s AI assistant is taking over the manual and repetitive tasks from legal professionals, AI technology will assist practitioners across various industries. Within a few years, using AI as a part of the workflow will feel completely normal. That is what we are also referring to with our slogan “DeepJudge: The future of the legal profession” – we envision that every lawyer or legal professional will be empowered by an AI in the not-too-distant future! What is something you wish you knew when you started putting together your business? That it is ok to not have everything figured out at the beginning and to ask for help as much as possible. When we started, we applied for various grant applications that required extensive documentation, financial projections and exact strategy. It is easy to get overwhelmed, but we learned that it is ok to shape and adapt things along the way. As we get more and more feedback from our customers, the takeaways are also reflected in our strategy and planning.

Another thing is to not reinvent the wheel, but to leverage the experience of investors, mentors, advisors and other startups. The ETH spin-off and the Swiss startup ecosystem in general are very friendly and open. We saved a lot of time and worries asking for advice on how they have handled similar issues we are dealing with at the moment.

What is the most challenging part of having a startup like DeepJudge AG? All of our tools are built on top of our context-aware semantic AI that forms a notion of understanding of the content of legal documents. Understanding the content of legal documents and how they relate to all other documents is really the key to automating a lot of manual and repetitive processes. And in the legal profession, there is no shortage of such tasks. Hence, we often get requests to solve a particular customer problem, which can be very tempting. This is particularly amplified given our research background and us having a knack for interesting problems. Yet, we have to focus on our mission and build the products first that we believe to have the biggest impact and are scalable internationally.


bottom of page