Evidence-Grounded Agents for Oncology Prior Authorization

AI Engineer16m 41s
0 comments · 0 votesOpen discussionClose discussion
Sign in to join the discussion

    Video summary

    Anant Shankhdhar describes Risa's architecture for oncology prior authorization as a healthcare-administration workflow rather than a replacement for clinical judgement. Four agent responsibilities cover insurance eligibility, authorization requirements, medical-necessity evidence and submission, with deterministic coverage checks deciding whether an order can proceed.

    Anant Shankhdhar explains why extracting a status from one note is insufficient. Existing authorization letters, payer portals and a period-specific SQL rule history provide corroborating evidence. Portal APIs and automation are normalized before decision-making, while model-generated configurations are used to repair some brittle portal interactions.

    For medical-necessity assessment, Anant Shankhdhar describes linking patient facts, clinical notes and payer criteria, including supporting and contradictory evidence. Confidence and unresolved questions determine escalation to clinicians. The talk presents an evolving conditional-automation system, not measured universal accuracy or a claim that every oncology workflow can run without people.

    Original YouTube thumbnailWatch on YouTube

    Share this page

    Anant Shankhdhar in a blue shirt against black beside the blue-and-white headline EVIDENCE BEFORE ACTION. Framed in blue with WWW.ARTIFICIAL-INTELLIGENCE.VIDEO, 20 July 2026 and duration 16m 41s.

    Risa's prior-authorization architecture separates insurance checks, evidence gathering and submission, keeping uncertain or contradictory cases subject to human review.