DeepJudge, founded by ex-Googlers and supported by former executives from successful legal tech companies is announcing first-of-its-kind Knowledge Search, a solution that enables legal professionals to search their entire document management systems using natural language.
“Across the globe, law firms and legal departments have accumulated hundreds of millions of documents containing decades of invaluable expertise, but they wrestle with error-prone keyword searches to surface relevant information,” shared Paulina Grnarova, co-founder and CEO of DeepJudge, who was named a Digital Shaper in Switzerland and appeared on the prestigious Forbes 30 Under 30 list. “That’s where we come in. We connect legal professionals to their internal documents, eliminating the hassle of finding and creating time for higher-value tasks such as strategy development, client engagement and more, ultimately increasing their productivity and revenue generation.”
The founding team of DeepJudge comprises Paulina Grnarova, Yannic Kilcher and Kevin Roth, all ex-Googlers and PhDs in Artificial Intelligence from ETH Zurich. They’re supported by seasoned advisors Jan Puzicha, who founded and scaled Recommind (now OpenText) as CTO, and Steve Obenski, whose strategic leadership propelled Kira Systems’ growth from its early stages through successful exit. Kennan Samman, who previously led global product sales at Litera has been appointed Chief Revenue Officer to spearhead DeepJudge’s expansion initiatives globally. Kennan Samman is a vision-driven growth strategist whose exceptional sales, leadership and client relationship management skills have been demonstrated throughout his career at prominent legal technology companies like Litera, Kira Systems and Thomson Reuters. This combination of experienced legal tech advisors and breakthrough technology researchers puts DeepJudge in a unique position to deliver high-impact solutions to the global market.
DeepJudge’s Knowledge Search is powered by the company’s cutting-edge and proprietary Colinear Technology, which blends the capabilities of semantic and keyword search to deliver an effortless knowledge retrieval experience. Colinear Technology uses advanced Artificial Intelligence models to understand the content within each document regardless of document type, and enables legal professionals to search in their own words, without struggling to find the perfect keywords. Additionally, each document is automatically classified into an extensive taxonomy for easy filtering and AI-predicted tags enrich existing metadata. With DeepJudge’s Knowledge Search, legal professionals can efficiently reuse prior work product to provide better client service.
Unlike some AI technology, DeepJudge’s Knowledge Search ensures that no information is shared with external third parties, respecting existing security protocols whether deployed on-premise or in a secure cloud environment. The tool automatically integrates to the DMS via a secure single sign-on, including iManage Work and Microsoft SharePoint, maintaining all native ethical walls and access permissions. Knowledge Search is 100% ready to use out-of-the-box with no additional training, and its performance can be optionally fine-tuned on client and enterprise data.
“We’ve built a truly robust solution using the latest research on Large Language Models (LLMs). With the ability to scale to hundreds of millions of documents while ensuring absolute privacy for sensitive data, we created an intuitive search experience that integrates effortlessly into existing systems,” shared Yannic Kilcher, DeepJudge co-founder and CTO, who also runs the biggest YouTube channel on Machine Learning Research and played a leading role in the development of OpenAssistant, an open source alternative to ChatGPT.
With Knowledge Search, DeepJudge forms a strong foundation for deploying trusted generative AI models. The ability to surface the most relevant documents from a firm’s document collection allows generative AI models to ground all generated data in factual enterprise information, significantly reducing the risk of hallucinations. Furthermore, by understanding the intent and contextual meaning of search queries, Knowledge Search can counteract a generative model’s reliance on purely statistical parameters, improving the quality of generated output and enabling instant verification and citing of source data.
DeepJudge has developed a strong presence in Switzerland since its founding in 2021 and has garnered the support of all major awards and grants in the country. The company counts top law firms among its customers and is backed by prominent investors such as Dr. Felix Ehrat, ex-General Counsel of Novartis, Daniel Sauter, ex-Chairman at Julius Bär, Dr. Christian Wenger, partner at Wenger & Vieli, Dr. Thomas Hofmann, former Director of Google and Professor in AI at ETH Zurich, and Dr. Thomas Duebendorfer, serial entrepreneur and president of SICTIC.
