Homburger Puts Institutional Knowledge at the Center of Its AI Strategy
Homburger is one of Switzerland’s leading business law firms, advising companies and entrepreneurs on complex transactions, disputes, regulatory matters, and other business-critical legal issues.
As AI changes how legal work gets done, Homburger is looking beyond access to general-purpose AI capabilities. The firm is focused on combining those capabilities with something uniquely its own: the knowledge and experience accumulated through years of legal work.
“We have at our fingertips the entire knowledge of what is market practice, for example, in a particular area.”
The Challenge
Homburger’s lawyers have built up years of knowledge and experience across millions of documents in the firm’s document management system. That prior work can provide valuable context for new matters—from understanding market practice to seeing how the firm has approached a similar issue before.
Historically, finding that knowledge often depended on one of two things: knowing the exact words used in a document or knowing the person who had the answer.
Keyword search could work when lawyers knew precisely what they were looking for. But when they did not know the terminology used in a prior matter, relevant knowledge was effectively out of reach. The alternative was often to identify and ask colleagues who might have the right experience.
“If you're not really sure what exact specific words have been used in a matter, then keyword search stops working. [Y]ou need to chase down the person internally who has the knowledge that you're looking for, which can take a long time.”
As generative AI became increasingly capable, Homburger saw a second challenge and opportunity. Anyone could access the same general-purpose models. The opportunity was to give those models context from Homburger’s own work.
Homburger wanted to make its collective knowledge easy to access and apply across the firm:
- Find Knowledge Without Exact Keywords: Enable lawyers to find relevant prior work using natural-language questions, even when their terminology does not match the source documents
- Reduce Reliance on Individual Knowledge: Make experience available without requiring lawyers to know which colleague or team has handled a similar issue before.
- Bring Firm Context to AI: Give AI access to relevant examples and context from Homburger’s own prior work.
- Maintain Trust and Control: Enable lawyers to trace AI-generated outputs back to the underlying sources and verify the quality of the result.
The Solution
DeepJudge turns Homburger’s knowledge and experience into institutional intelligence that lawyers and AI can put to work.
Connected directly to Homburger’s document management system, DeepJudge enables lawyers to work with the firm’s existing knowledge without moving or manually curating it. Lawyers can search across prior work in natural language, find relevant information even when different terminology was used, and bring that context into AI-assisted workflows.
For Homburger, this becomes increasingly important as AI takes on a larger role in legal work. General-purpose models can produce sophisticated outputs, but they do not inherently know how Homburger has handled similar matters, what its lawyers have negotiated before, or what experience the firm has accrued over time.
How Homburger Uses DeepJudge
Understand Market Practice from Prior Work
Lawyers can use DeepJudge to look find and analyze prior work across the firm and understand how particular provisions, issues, or deal terms have been handled in practice.
Rather than relying on exact keywords, lawyers can formulate queries in natural language and surface relevant information even when the underlying documents use different terminology.
That gives lawyers access to a broader body of firm experience when advising on a new matter.
“Even if they don't match the words exactly, you're still able to surface the relevant information that you need.”
Bring Firm Context into AI-Powered Work
As advanced AI becomes broadly available, Homburger appreciates the critical distinction between what a general-purpose model can produce and what it can produce with context from the firm’s own work.
A model may be able to generate a share purchase agreement on its own. But without access to Homburger’s knowledge, that agreement is based on the same generic information available to anyone using the model.
DeepJudge enables Homburger to bring examples and context from its own prior work into those AI workflows.
“Anybody can log in to a large language model platform. They can ask the model to generate an SPA and the model can generate probably a pretty decent SPA. But it'll be generic, and everybody else who uses that platform will also have access to the exact same SPA.”
Start with Analysis, Informed by Context
Rather than beginning with the mechanics of drafting, lawyers can start by bringing together information about the matter, the client, relevant prior work, and market practice. AI can help analyze that context and support a draft tailored to the specific situation.
That shifts more of the lawyer’s attention toward the judgment required to understand what is different about the matter and how the work should respond.
“The first thing is … all the relevant information we have on that particular project. Get an analysis: What is special, for example, for that particular client? How would we have to adjust the SPA? What is the market practice?”
Verify AI Outputs Against Their Sources
As AI takes on more of the drafting and analysis process, Homburger sees critical thinking as increasingly important.
Lawyers need to understand the output they receive, identify the sources informing it, and determine whether the result meets the quality expected of the firm.
DeepJudge makes the underlying firm knowledge visible so lawyers can trace answers back to their sources rather than treating generated output as a black box.
“We need to be able to ensure that all the outputs that we generate, that we can verify them and that the quality is up to Homburger standards.”
The Results
DeepJudge has made Homburger’s cumulative knowledge easier to find and apply while giving the firm a way to bring that knowledge into a new generation of AI-enabled legal work.
1. Firm Knowledge at Lawyers’ Fingertips
Lawyers no longer need to know the exact terminology used in a prior document—or which colleague might know where to find it—to benefit from Homburger’s prior experience.
With more than 80% of the firm’s legal professionals regularly using DeepJudge, institutional knowledge is becoming part of lawyers’ everyday work.
DeepJudge gives them direct access to relevant knowledge across the firm’s document collection, making prior work available when and where it is needed.
2. More Time for High-Value Legal Work
Reducing the time spent searching for information or tracking down colleagues gives lawyers more time to focus on analysis, judgment, and client work.
“Since you're not running around talking to five different partners about where the correct data is, it leaves a lot more time for you to focus on high-value work.”
3. AI Informed by What Makes Homburger Different
General-purpose AI capabilities are increasingly accessible to law firms, clients, and competitors alike.
Homburger’s own prior work, experience, and understanding of market practice are not.
By making that knowledge available to AI, Homburger's lawyers can move beyond generic outputs and bring firm-specific context into their work.
4. A New Model for Legal Work
Homburger expects AI to change the substance of lawyers’ work, not simply make existing processes faster.
As tools take on more drafting and initial analysis, the lawyer’s ability to evaluate an output, understand its sources, identify what is distinctive about a client or matter, and exercise critical judgment becomes even more important.
“What is more important for the lawyers who use these tools is critical thinking. I think that's been my mantra for a long time. What they need to learn is to understand the output that they get.”
Built Through Collaboration
Homburger’s work with DeepJudge has continued to evolve alongside the firm’s approach to AI.
For Homburger, that relationship is not simply about adopting a technology product. The firm provides feedback based on how its lawyers work and what they need, and that feedback is reflected in how DeepJudge develops the product.
“You really feel that the feedback you provide to DeepJudge makes a difference in how the product works. That made a huge impression on us, and that's definitely one of the main reasons why we continue to work with DeepJudge.”
As AI becomes part of more legal workflows, Homburger is building on an advantage that general-purpose models cannot provide on their own: the knowledge, experience, and judgment accumulated through the firm’s own work.
DeepJudge turns that institutional knowledge into intelligence Homburger can put to work—helping its lawyers build on the firm’s experience as legal work evolves.