What you get every issue: Three curated healthcare AI stories with my take. Companies doing the work. Skills and techniques worth knowing. And opportunities worth pursuing. No hype. Just signal.

Three things worth your time this week:

  1. PRIOR AUTH IS NOW AN 18-SECOND EVENT
    Sure Scripts just announced its PA Automation solution reached 68,000 prescribers across 42 health systems — a 50% increase since December. Median approval time when all criteria are met: 18 seconds.
    The headline: AI is finally killing the fax machine.
    The reality: 34% automated approval rate means 66% still require human review. The hard part isn't the approvals — it's the exceptions.
    My take: When 18 seconds becomes the benchmark, every health system that still takes 7 days looks negligent. That gap will drive procurement decisions in 2026. We'll see how this one pans out.
    Read it: surescripts.com
  2. CMS QUIETLY CHANGED THE RPM RULES AND ALMOST NO ONE NOTICED
    The 2026 Physician Fee Schedule added two new CPT codes that let providers bill for remote monitoring with as few as 2 days of data — down from 16. This is the biggest RPM reimbursement expansion since the original codes were created.
    Why it matters: Short-term monitoring — post-discharge, medication adjustments, transitional care — was previously unbillable. Now it is. The patient population eligible for RPM just got significantly larger.
    The pattern: CMS keeps quietly expanding the reimbursement envelope for remote monitoring. Every year, the clinical and financial case for RPM platforms gets stronger.
    My take: This is a direct tailwind for monitoring programs. Companies that are not building RPM capacity now are falling behind.
    Full story: intuitionlabs.ai/articles/ai-remote-patient-monitoring
  3. LESS THAN 1% OF LLM HEALTHCARE STUDIES HAPPENED IN REAL CLINICAL SETTINGS
    Nature Medicine published a meta-analysis: fewer than 1% of studies on large language models in healthcare were conducted in live clinical settings. Most evidence comes from simulated or synthetic environments.
    But here's the nuance: This isn't an indictment of healthcare AI — it's a measurement problem. The tools are deploying faster than the research can follow.
    Why this matters for builders: The first company in any clinical AI category that publishes real-world outcome data from live deployments wins the credibility race. And right now, almost no one has.
    Source: telehealth.org

COMPANIES DOING THE WORK
Sure Sscripts — Building the network layer for real-time PA automation across 42 health systems.
Abridge — Scaling ambient AI to 80M patient conversations in 2026 while simultaneously wiring prior auth through Availity.
Doximity — Betting its 85% physician network penetration on becoming the AI platform doctors actually use — not just the one they're signed up for.

AI TOOLS WORTH KNOWING
Dragon Copilot (Microsoft) — Ambient AI documentation now used in 50%+ of encounters at major health systems. Deepest Epic integration in the market.
LunaBill — AI voice callers for insurance claim follow-up. 10x more claims resolved per biller. $764K ARR. UC Health and Mayo are customers.
Suki AI — Ambient AI with the most EHR integrations (Epic, Oracle, athenahealth, MEDITECH). Now embedding in health plan care management via HealthEdge.

WORK & PARTNERSHIP OPPORTUNITIES
CPO / Co-Founder — Damo Consulting health AI intelligence platform. Enterprise clients. Wants SaaS and information products background.
Chief AI Officer — Health systems standing up governance roles now. MD/DO with informatics background strongly preferred.
Director of Referral Operations — Home health agencies hiring to run AI-assisted intake programs.

WHAT I'M WORKING ON:
The Health AI report. Experimenting with new and old tools including: the Claude API. Open Codex, Hermes and Composio a super connector.

— Victor

Reply to this email if you want to discuss any of these. I read every message.

Yuuki Edge • Healthcare AI Intelligence • yuukiedge.com