Saffron Health: AI-powered care coordination for primary care
AI agents that fully automate specialist referrals
Primary care providers make over 100 million referrals to specialists every year, leading to tens of millions of hours in manual coordination work. We built AI agents to automate the entire specialist referral process for primary care groups.
Problem
Managing specialist referrals is one of the most time-consuming workflows in primary care. Staff must manually copy data between systems, complete prior authorizations, call specialists to verify information and request medical records, and coordinate with patients.
Many clinics are forced to hire or outsource their referrals, but they’re still understaffed. As a result, referrals often never get sent, are returned for missing authorizations, or go to out-of-network providers. Nonetheless, poor referral management has significant consequences:
- In value-based care models, specialist routing has a massive impact on medical spending. Out-of-network costs and unnecessary procedures are major cost drivers.
- Groups that depend on federal funding or grants, like FQHCs, can lose funding due to poor referral follow-through.
- Patients who experience delays in care because of referrals are likely to churn and find a new primary care provider.
Product
We partnered with a network of rural health clinics in west Texas to understand the specialist referral process. We started by manually handling referrals, then built AI agents to automate the whole workflow. Our agents are live across the network, where we’ve:
- Sped up referral processing time by 82%
- Found available appointment times that were 55% sooner
- Led to 44% fewer referral kickbacks
Team
Tanishq Kancharla has worked in startups throughout his career. He was a product engineer at Shortwave, leading development on their iOS and Desktop apps, used by tens of thousands of people. Prior to this he double majored in physics and computer science at Carnegie Mellon.
Michael Kronovet was technical lead overseeing Palantir's work with the US State Department, quadrupling annual revenue without increasing headcount. Previously, he built causal inference models at a healthcare startup. He graduated in 3 years from Carnegie Mellon with a BS in statistics and machine learning.
Our Ask
If you know anyone who works in or adjacent to primary care we’d love to be connected!
- Some examples: people who work in MSOs, ACOs, primary care groups, RHCs, FQHCs or payor/insurance groups.