Fathom + CVS Health: Why Payers Are Betting on AI Coding
CVS Health Ventures invested in Fathom in May 2026. Here’s what a major payer backing autonomous medical coding means for RCM teams and coders.
·
-
-
Ensemble and Cohere Build First RCM-Native LLM for Coders
Generic LLMs weren’t built for RCM. Ensemble and Cohere are building the first revenue cycle-native LLM, expected H2 2026—and it changes how coders should evaluate AI.
·
-
R1 Phare OS: Medical Coding Meets Real-Time Adjudication
R1 launched Payer Atlas and Phare Intelligence on June 18, 2026—two capabilities designed to move medical coding toward real-time claim adjudication.
·
-
Omega Healthcare’s PEAK Matrix Win: What It Means for RCM AI
Omega Healthcare is the only provider to earn both Leader and Star Performer in Everest Group’s 2026 RCM PEAK Matrix — a signal about agentic AI in the revenue cycle.
·
-
AGS Health Debuts Agentic AI for Medical Coding, RCM
AGS Health’s new InnovationWorks center applies agentic AI to autonomous coding, CAC, and CDI via AWS Bedrock AgentCore — built for auditability, not speed.
·
-
athenahealth Express Coding: Ambulatory Coding Arrives
athenahealth’s Express Coding automates 51% of ambulatory charges and routes one-third of claims to billing with zero coder review. Here’s what it means.
·
-
AI Governance in RCM: What Black Book’s 2026 Data Says
Black Book’s 2026 RCM survey found 63% of hospitals demand auditable AI—and 69% require human oversight before AI takes any coding or appeal action.
·
-
Mayo-Microsoft Healthcare AI Model: What Coders Must Know
Mayo Clinic and Microsoft are building a frontier AI model for healthcare—trained on de-identified clinical data. Here’s what medical coders need to understand about this shift.
·
-
Carlyle’s AI-Native RCM Deal: What Medical Coders Need to Know
Carlyle’s acquisition of Knack RCM and EqualizeRCM signals a major shift toward agentic AI in revenue cycle management.
·
-
The Wrapper Problem: Why Medical Coding AI Needs Its Own Language Model
Most medical coding AI is just a general LLM with clever prompts — and that’s the problem. Here’s what purpose-built RCM models do differently.
·