Key results
Result highlights
- $100K–$250K annual savings via reduced attorney hours
The story
One of the world's largest biotech companies requires exhaustive legal research across roughly 50,000 cases from the European Patent Office to defend its intellectual property.
Attorneys spent hours on manual keyword searches that often missed relevant precedents, while a subsequent rollout of isolated AI tools created a fragmented and confusing workflow.
The legal team developed an agentic AI interface that acts as an orchestrator, automatically routing queries to specialized sub-agents for case law, deep search, and internal documents. This system unifies fragmented tools into a single chat-based workspace where attorneys can summarize documents and verify statements without switching contexts. The implementation utilizes a low-code platform, enabling non-technical staff to build and refine workflows without IT bottlenecks.
Quotes
“Evaluating different prompts and estimating token costs using/comparing different LLMs was easy on the Dataiku platform.”
“The entry barrier to starting with AI and data analytics tools is remarkably low when using the Dataiku platform, making it accessible even for non-coders.”
“Using the Dataiku platform and leveraging self-service or citizen development, one can achieve rapid proof-of-concept or even proof-of-value without the need for additional IT experts.”
The company
Roche
roche.comBiopharmaceutical and diagnostic solutions for healthcare and drug development.
Scope & timeline
- New GenAI project build time cut from months to days