Thrilled to announce that after a year of research and development, our

Thrilled to announce that after a year of research and development, our conversational visual analytics dashboard for single-cell data is now live! 🎉

This feature transforms how scientists interact with complex biological datasets - simply chat with your data to quickly uncover insights without needing to master the UI first.

Working with the wonderful Jane Adams, we tackled a fundamental challenge in life sciences visualization: how do we make complex visual analytics dashboards more accessible from a data exploration perspective?

Our solution is a mixed-initiative system where:

• An LLM-powered agent reconfigures the dashboard in response to natural language queries
• Users maintain complete control and can seamlessly switch between conversation and direct manipulation
• The agent operates within the same constraints as human users, ensuring all responses are verifiable and grounded in Ozette’s high-quality and standardized analysis pipelines

Unlike a classic chat interface, our AI agent doesn’t just generate text - it drives actual dashboard reconfigurations that show you the evidence behind its answers through interactive and linked visualizations.

This bridges the gap between conversation and visual data exploration, dramatically reducing the learning curve for complex scientific dashboards.

Huge thanks to the entire Ozette team and especially to Jane Adams who developed the initial prototype during her internship with us last year. 👏

We’ll share more details in the coming weeks/months.