Ember

Who are the best up & coming engineers (in school, just graduated, dropped out)? Tag anyone you know who's cracked at understanding complex workflows and has built enterprise products that scale. I'm looking for an intern / apprentice to work closely with me. We want: 1. The project you're most proud of 2. A direct reference

A number of our RCM partners are actively looking to hire medical billers who are comfortable with tech and excited to work with modern tools. If you know someone who fits the bill (pun intended 😎), tag them below or DM me their profile — happy to make introductions. Let’s help great talent land in the right seats. 🙌

The "Epic-first" strategy in healthcare IT is a fascinating dance I've been watching unfold. Most health systems start their vendor conversations the same way: "What does Epic have for this?" They test it, hope it works, then scramble to fill gaps. It's become the default playbook — and honestly, it's steering the entire vendor ecosystem. The shift in AI sentiment from last year to now? Night and day. Before: Fear. Skepticism. Endless committees debating the "risks of AI." Now: Pilots everywhere. Real implementations. Actual results. Once teams get their hands on these tools, reality sets in. AI automation platforms aren't magic wands — they're more like skilled interns who need training. Just like scribes don't automatically match each provider's documentation style, AI needs change management to drive real adoption. And of course, the eternal revenue cycle hits remain: denial appeals, prior auth, patient access. But the approach has fundamentally shifted from "how do we fight denials" to "how do we prevent them through better analytics and deliver hard ROI." Patrick McDermott from Geisinger laid out the brutal math: - Hospital margins: 2-3% - Payer margins: 20-30% - $3T in healthcare waste Other warnings about AI implementation stuck with me: "Don't automate bad processes." The current payer-hospital arms race is already broken. Slapping AI on top just makes it fail faster. The winning formula? Redesign first, automate second. Focus on eliminating waste, not accelerating inefficient workflows. The vendors who'll win aren't those chasing Epic's shadow — they're the ones solving the high-pain problems Epic won't touch yet, with tech that actually fits into clinical workflows.

🏥 Day 3 at Becker's Healthcare IT & RCM Conference — and the energy is electric. A few sessions for me that hit differently: 1️⃣ Katy Howard shared one stat that stopped the room: $33M lost to auth delays. Cancer patients waiting 14 days for chemo authorization? That's broken. It's clear we need to put patients at the center of every process conversation. Show physicians the revenue impact to drive buy-in. Their centralization play (240 clinics!) proves you can scale without proportional staffing when people, processes, and technology actually align. 2️⃣ Healthcare systems averaging ~$1M quarterly from AI automation. The priority? Reducing mental load on denials and high-skill tasks. Mike Laukaitis's "RCM 3.0" vision caught my attention — training next-gen leaders on dashboard analysis and agentic AI oversight. AI extends skills, doesn't replace them. Launch timeline that shocked me: 1-2 months for 4,000+ person teams. 3️⃣ Change Healthcare breach = wake-up call. Tanvir Singh Sahsi, MD, MHA made it clear: establish safety and backup systems before chasing AI innovations. Real-time eligibility verification delivers the biggest ROI, freeing staff from screens to focus on patients. The upskilling focus surprised me — soft skills over technical. Adaptability, communication, and the deep expertise to catch AI hallucinations. 4️⃣ Emory's OR authorization workflow fix saved $1.5M in month one. How? Jamie Watson walked us through improved post-op to front-end communication. Their golden rule: refuse to make it finance-only. Physician-led, physician-owned initiatives + transparent data sharing = trust and results. Their Z-code hard stop eliminated ALL medical necessity denials for that category. Grateful to be part of this community balancing innovation with humanity.

Epic engineers gave us their playbook: See an existing button? A form to fill? That's Epic's territory — they're already building it out internally. Smart vendors should look elsewhere. Like imaging. Epic doesn't touch radiology image processing (yet). That's open field. But even when Epic does chase a workflow, results can be... underwhelming. They poured resources into a sepsis prediction model they still champion, but JAMA found it performs barely better than a coin flip (article linked in comments). Meanwhile, Epic's opening APIs from the Abridge workshop to more vendors. They're watching the ecosystem build what they won't. My take? Skip the crowded lanes. Build for high-acuity, high-pain scenarios where humans struggle most. The hardest problems are where real value lives.