DeepJudge Introduces AI Workflows, Connecting Gen AI To Enterprise Search of Law Firms’ Internal Data

DeepJudge, developer of an AI-powered search platform for internal law firm data, has launched AI Workflows, a new feature of its product that allows law firms to deploy AI agents and applications across their internal firm data. The feature provides built-in AI workflows and also allows firms to create custom AI applications via a low-code/no-code interface.

The feature, the Swiss-based company says, marks a major leap forward in legal AI by enabling law firms to build AI-powered workflows that are tailored to their specific needs — without requiring firms to restructure or upload their data.

The company said that several prominent customers are already implementing the platform, including the law firm Gunderson Dettmer and CMS Switzerland. Gunderson, which made news in 2023 as the first U.S. law firm to develop an proprietary internal gen AI app, ChatGD, is adopting DeepJudge as a replacement for that app.

Founded by ex-Google researchers with PhDs in AI from ETH Zurich, the company was already known for its powerful enterprise search engine, which can ingest and index unstructured documents from a firm’s DMS, Microsoft ecosystem, client portals, and more, and make it all accessible through a single search.

The new AI Workflows build on top of that search product, enabling firms to deploy and develop applications that leverage their own institutional knowledge, using either pre-built workflows or through a no-code editor.

The company says it is the first solution that provides both enterprise-grade search capabilities and the ability to create AI-powered workflows tailored to legal professionals’ specific needs.

‘AI Should Work for You’

“AI should work for you — not make you work for it,” Paulina Grnarova, cofounder and CEO of DeepJudge, said in a press release. “Too many vendors require firms to ‘clean up their data’ before AI delivers value. That’s not solving the problem, it’s shifting the burden. But legal teams don’t work with perfectly structured information, and AI shouldn’t require it either.”

“Instead of hunting for documents to upload to an AI platform, everything you need is already there,” Joe Green, chief innovation officer at Gunderson Dettmer, said in a statement provided by DeepJudge. “What excites me most is how DeepJudge search powers our AI strategy. Fast, efficient, and compliant access to the right information is the foundation for AI applications that help us do more with our knowledge base.”

The AI Workflows feature includes both pre-built workflows and a workflow builder that enables firms to create customized solutions. Built-in workflows include functions like summarizing documents, creating client or matter overviews by aggregating relevant information, and analyzing clause variations across multiple documents. More complex custom workflows can be created through DeepJudge’s no-code editor.

Data Trapped In Silos

Yannic Kilcher, CTO of DeepJudge, said that the company developed the platform to address fundamental challenges of enterprise search in legal settings. “Much of the work in building AI apps is handling the data integration, data governance and finding the right information for any query,” he said. “Legal data is trapped in silos — legacy databases, document management systems, SaaS platforms, and collaboration tools.”

DeepJudge’s platform attempts to solve these problems by providing seamless integration with firm systems while respecting existing security and access permissions. This approach makes it possible to scale AI adoption across an enterprise while maintaining appropriate data governance, Kilcher said in a briefing yesterday.

The company says its AI Workflows feature is particularly valuable because it pairs powerful search capabilities that scale to millions of documents with large language models that can then be used to work with the data found through search.

“LLMs aren’t so good at reading hundreds of millions of documents,” Kilcher said in our briefing. “But when you pair it with the search engine, which is really good at looking at 100 million documents and giving you the 20 that are most relevant to what you want to know, then the LLM can read the 20 and give you, in a more nuanced way, the answer to your question.”

A simple example Kilcher and Grnarova gave me is that of an attorney about to get on a call with a client. The attorney can use the search to find everything that the firm has done for the client, and then use the AI to take those results and generate a summary and timeline.

Kilcher said the search and LLMs could also be combined in a series of prompts. As an example, an attorney with a new case might search for all the documents the firm has that are similar to that case. Having found those documents, the attorney could then prompt the platform to summarize the context around all those found documents.

“Workflows are essentially a platform,” Kilcher said. “You can implement single-prompt interactions, you can implement entire agents, you can implement multi-step, multi-LLM, multi retrieval things. It’s a dynamic platform that allows you to implement all these different types of use cases.”

Nick Zaugg, dispute resolution partner at CMS Switzerland, emphasized the platform’s ability to leverage existing knowledge: “What sets DeepJudge apart is how quick and painless it is to leverage your firm’s own knowledge base. Clients come to us for our past expertise, built up over years. That expertise lives in our documents, and so they are the essential starting point for our AI-driven client work product.”

Over the past year, the company has expanded its presence in the U.S. market, CEO Grnarova told me.

If you are interested in learning more, DeepJudge will showcase AI Workflows in a live webinar featuring Gunderson Dettmer on March 18, 2025. You can register at: https://deepjudge.ai.