AI in Healthcare: Model Speed, Data Quality, and Governance

Hetal Patel · VP, Healthcare Economics and Data Science · HarmonyCares

Nov 2026AI Data

01 — The conversation

GuestHetal PatelVP of Healthcare Economics and Data ScienceHarmonyCaresProfile →

AI is helping healthcare teams build predictive models faster, reduce administrative work, and improve analytics delivery. Hetal Patel, VP of Healthcare Economics and Data Science at HarmonyCares, explains why human judgment, bias testing, and clinical adoption remain essential even when model development time drops by more than 50%.

Key ideas

01AI scribes and contract analysis agents reduce administrative work for clinicians and business teams.

02Hetal's team shortened development time for a readmissions model by more than 50%.

03Faster model delivery increases pressure on clinicians to evaluate and adopt new tools.

04Model accuracy alone is not enough without bias testing, clinical validation, and human oversight.

05Better access to clinical notes and patient records may strengthen predictive models while raising privacy concerns.

Healthcare organizations are finding practical ways to use AI, from reducing documentation work to accelerating data science projects. Hetal discusses how her team uses AI across analytics, code generation, contract interpretation, and data migration.

One example involves a patient readmissions model whose development time was reduced by more than 50%. Hetal explains why that improvement creates new pressure for clinicians to evaluate and adopt the resulting tools. She emphasizes that faster deployment is not always safer in healthcare.

The conversation explores how biased data can produce unfair outcomes, why checking and validating AI output can take more time than building models, and how clinical notes and better interoperability could improve patient predictions while raising privacy questions.