Assort Health secures $120M to expand AI agents across patient journeys

Assort Health is arguably one of the first at-scale demonstrations of agentic AI operating in the real economy, not just in demos. Automating scheduling, intake, referrals and payments across hundreds of millions of patient interactions is a qualitatively different thing from answering questions in a chatbox. It means multi‑step reasoning, integration with legacy systems and non‑trivial error budgets that CFOs and compliance officers have signed off on.

From an AGI perspective, this shows how quickly “narrow” agents can colonise complex service domains once they’re given a focused remit and high-quality domain data. The capital going into Assort is a bet that healthcare administration is a beachhead: if you can reliably orchestrate agents across that maze of edge cases, moving into adjacent workflows—prior authorisations, revenue cycle, even some clinical support—becomes more plausible. That creates a virtuous cycle: more agent usage → richer datasets on human/AI decision patterns → better models and policies.

It also hints at a strategic divide between frontier labs and application-layer winners. Even if hospitals are ultimately running OpenAI, Anthropic or Google models under the hood, the company that owns the workflows and data (Assort in this case) captures much of the value and practical influence over how capable systems are allowed to behave.