Finding the right medical dataset should start with describing the data you need, not figuring out which combination of filters will get you there. Today we're launching AI Search Assistant in Atlas, designed to help medical AI developers find and build datasets faster using natural language. Instead of manually translating dataset requirements into individual search filters, Atlas users can now describe the cohort they need in plain language. The AI Search Assistant interprets the request and automatically applies the relevant filters.
The AI Search Assistant is a natural-language search feature built into our Atlas platform. For example, rather than manually selecting criteria for modality, anatomy, demographics, equipment, or other available data fields, a developer can describe those requirements in a single prompt. Atlas then translates that request into structured search criteria, allowing the user to see exactly how the cohort has been defined and refine it further if needed. This combines the speed of conversational AI with the precision and transparency of structured medical data search.
During model development, teams may need to narrow a cohort by patient demographics, include or exclude specific manufacturers, adjust clinical criteria, or explore whether enough relevant cases are available. Traditionally, each change can mean returning to a set of filters and working out how the updated requirement maps to the search interface. With AI Search Assistant, refinement can be much more conversational. Developers can start with what they need, review the resulting filters, then continue adjusting the cohort as their requirements evolve.
The result is less time spent navigating search logic and more time evaluating whether the data is right for the task.
AI Search Assistant does not replace Atlas's structured search capabilities. It provides a faster way to use them. The assistant converts natural-language requests into Atlas filters that users can review, edit, and refine.
AI Search Assistant is now available to all Atlas users.
Existing users can log in and start searching with natural language today.
For medical AI teams looking for a faster way to discover and build high-quality medical datasets, explore Atlas and see how Gradient Health is simplifying medical data access.
