Pelica Health

Health plans, IPAs, ACOs, and medical groups managing risk-bearing contracts have their data scattered across claims, EHR, pharmacy, and ADT systems, with risk adjustment, quality, and care teams each working from separate files or spreadsheets. This causes duplicate outreach, such as the same patient being called three times by different teams, sometimes in a language they don't speak. Submission windows for HCC capture close before suspect codes surface because data arrives too late or in disconnected silos. One customer, a VP of Quality, describes building a strong predictive model but being unable to get it into a provider's hands, making the analysis useless in practice.

Pelica provides a canonical record built from live ingestion across claims (MMR, MAO-004, MOR, 277CA), EHR (FHIR/HL7), pharmacy (PDE), labs, ADT, and payer SFTP drops, described as built for ECDS volumes at 35 to 75 times the supplemental data of hybrid sampling. On top of this sits an intelligence layer of six copilots, covering risk adjustment, quality and Stars, pharmacy and Part D, provider network, care management, and a cross-cutting AI Data Analyst, each powered by what the page calls 80+ specialized AI agents, with every recommendation citing its evidence and built-in trumping logic and RADV-defensibility. An action layer then pushes outputs out through TCPA-compliant multilingual outbound voice, EMR overlays (Epic SMART-on-FHIR, Athena marketplace), a provider and staff portal, and a coder workspace. Because all copilots read and write to the same record, an HCC captured by the risk copilot can inform the provider network agenda, and a Part D gap closed by pharmacy updates the Stars projection.