Neo AI has exclusive access to two datasets: Practo's India-scale clinical network and Cent's head-to-toe, longitudinal cohort. We turn them into datasets, models and partnerships.
1.1M+ doctors and 680K+ hospitals and clinics on one platform, across search, booking, visits, treatment and follow-up.
Every member measured head to toe (MRI, CT, DEXA, echo, 100+ biomarkers and genetics), then tracked between scans.
Practo's labs, clinicians and patients give Cent ground truth. Cent's models flow back into every patient's record.
Practo covers the care system. Cent covers the person, with every modality on the same individual.
Roughly 10× more data every 18 months: a cohort at the scale UK Biobank took about a decade to build.
Genotype-to-phenotype models on whole genome and organ imaging from the same people.
Multimodal discovery across 100+ biomarkers, 10,000 next, linked to imaging and outcomes.
Progression models and digital-twin control arms on 4M patients with 4+ years of history; patient-to-trial matching across 680K+ facilities.
Treatment-pathway and outcome models on 43M repeat patients and 438K surgical patients.
Why us: a population under-represented in global genomic data, measured head to toe and followed over time.
268B clinical tokens across 100M records, plus 13.3B multimodal tokens heading to 1.4T by 2030 Target.
Physician-graded: AI drafts a surgical second opinion with every claim cited, a physician signs off, and each review becomes a label.
5 agents in production handle 50K patient calls and replies a week, measured on every interaction.
A longitudinal model of each member: organ age, disease trajectories and a digital twin.
Why us: India-scale clinical data and a whole-person cohort, with AI already in production.
Link what the genome predicts to what the scan shows, member by member.
1 in 40 members had a critical finding and 1 in 4 a major one. The full cancer, cardiac, neuro and metabolic report takes 2 hours.
Cent's body-fat model is within 0.8 kg of DEXA; other at-home devices are off by ~5 kg.
Whole-genome sequencing and 10,000-biomarker panels are next in scope.
Why us: one of the few datasets anywhere with imaging, blood, genome, wearables and clinical data on the same people.
Every re-scan checks the last prediction, and the model retrains: predict, then prevent.
~80% of each new agent is reused, so the next one ships faster.
Practo: AI labs and data partners.
Cent: molecular and genomics companies, multimodal AI labs.
In frontier model development, and in molecular diagnostics and biomarkers, to help pick the first biomarkers, models and partners.
Practo: AI data-partnerships lead.
Cent: AI researchers, and go-to-market talent for the US and Middle East.
Why now: India-scale clinical data and a whole-person cohort, with AI already in production.