Vision
Most AI for biology surfaces the middle of the distribution. We help biopharma teams find the overlooked opportunities hidden in biological data.
Biological R&D has no equivalent of Google search: even strong teams explore only a fraction of the available search space, whether they're testing novel hypotheses or looking for new market opportunities. As a result, they're slow to surface meaningful opportunities, or miss them entirely.
We built Urchin, an agentic biological search and reasoning platform, to help biopharma teams systematically explore biological search spaces to identify non-obvious, high-value opportunities, from indication selection and drug repurposing to target prioritization and program strategy.
How Urchin Works
- Urchin designs query-specific agentic research pipelines that autonomously search, integrate, and reason over the world’s diverse and fragmented biological data — from public literature to private, quantitative experimental results. It can aggregate even weak signals across a vast, multi-modal array of biological knowledge bases that no human team could cover at scale.
- Urchin builds structured, inspectable reasoning traces, not just chat outputs or reports. Every claim is grounded in evidence and fully citable, so your team can evaluate the strength of each opportunity before committing resources to act upon it.
- Urchin's knowledge base grows with your program. Teams build upon and extend persistent traces as projects evolve, enabling rapid iterative discovery and decision continuity from early discovery through IND.
Our Approach
Finding alpha
Our ethos is to surface what others miss. Most AI tools for biology are optimized for the center of the search space: the well-trodden, heavily-cited territory where competitive crowding is highest. Urchin is built to find opportunities outside that center: overlooked drug classes, underexplored indications, weak signals that build conviction in aggregate.
Customized collaboration
Currently, we work closely with customers as a collaborative service, deploying Urchin in specific projects to help them answer high-value scientific questions. We've helped teams identify novel therapeutic hypotheses, prioritize indications for emerging platform technologies, and surface repurposing opportunities that human experts missed.
Human-level depth, superhuman throughput
Human scientists create robust reasoning chains across diverse and fragmented evidence, but lack throughput. Urchin combines the nuance and depth of expert reasoning with the scale of parallelized, automated search, rapidly mapping entire search spaces that would otherwise take months.
Building conviction
We recognize information is not the same thing as insight, and having an insight is not the same as getting to conviction. We help clients build conviction and make the best decisions based on available evidence through interactive strategy sprints, supported by our technology.