Bond Health

Bond Health plugs into EHRs, surfaces every eligible patient for active trials, and handles voice/text prescreening automatically. Sites see 50%+ faster screening and 2-3x enrollment. The criteria that determine eligibility do not live in structured data. Diagnosis codes and labs produce long lists of maybes, but real gates like disease severity, prior therapy failures, and contraindications sit in free-text notes. LLMs changed what's possible, but most teams still treat discovery as the bottleneck. Sites drown in eligible-patient lists because coordinators have no time to chart-review and call hundreds of people. Bond Health does the work end to end: protocol parsing, EHR screening against notes, and voice/text prescreening, so a coordinator receives a short list of pre-qualified patients. Rishabh built the pre-LLM version of patient stratification at Genie Life Sciences and saw firsthand that structured data was never enough. He then went to Harvard Medical School's biomedical informatics program to work on this problem the moment LLMs caught up. Co-founder Sammy Mustafa is a Harvard Medical School graduate and Regeneron Science Talent Scholar who led AI clinical research and trial operations at Mass General Brigham. Together they bring both the engineering and clinical operations depth. Three shifts converged: (1) LLMs can finally read free-text clinical notes at scale, which wasn't possible when Rishabh built the pre-LLM version at Genie Life Sciences; (2) Cures Act information-blocking rules and FHIR APIs standardized EHR data access, making integration feasible; (3) Trial registrations grew 113% from 2010 to 2024, sites face chronic coordinator shortages, and FDA diversity expectations push sponsors beyond over-tapped academic centers. The market now needs exactly what only recently became buildable.