Healthcare + AI is happening!
Welcome back to Founder Mode!
Today's newsletter will pull back the curtain on something we've been working on for the past few months. Improving healthcare has been something JJ and I have given considerable thought to. Founding Pretty Good AI has enabled us to make significant progress on a huge problem. Starting with a Voice AI Automation solution for Athena Health.
This week on Founder Mode, Rebecca Shufeldt, Founder and CEO of Ignite Healthcare Solutions, joined us. She discussed the real work that happens after launch in complex healthcare environments. She has led EHR rollouts, revenue cycle improvements, and change management in the complex healthcare environment. Her take was sharp and straightforward: the magic happens after go-live—if you keep going.
Jason said it well: progress in healthcare isn’t just big breakthroughs. It’s steady work after launch. I agree. When Jason co-founded Heal, the win wasn’t the app. It was more time with patients. Tech works when it gives time back. Today, AI can do that if we use it the right way once the system is live.
Day Zero Mindset: Go-Live Is the Start, Not the End
Launch day doesn’t prove the workflow works. It’s the first real test.
- Plan 30 days of fixes. Put time on calendars before launch. Treat it like a sprint.
- Keep scope tight. Pick the top three friction points. Fix them fast.
- Make changes visible. Share what shipped each week so teams feel progress.
“Go live is not the finish line. Go live is when you actually get to do the testing.”
— Rebecca Shufeldt
AI at the Front Door: Answer on the First Ring
The phone is still the front door of care. Miss that call, and you may miss the patient.
AI voice agents can:
- Manage schedules & reschedules
- Give directions and many other FAQs
- Handle medication refill requests
- Create tasks or cases in the EHR
AI can operate 24/7 and scale up or down as needed. A few principles we designed our initial solution to solve:
- Triage first. Let AI handle simple calls. Escalate fast for complex ones.
- Cut hold times. Aim for “human within two minutes” on escalations.
- Document everything. Push notes and tasks straight into the EHR.
Our customers are seeing over half of inbound calls handled end-to-end by AI. Fully contained. Staff focus on higher-value work. Patients get help right away.
Work at the Top of License
Doctors should do doctor work. Nurses should do nurse work. The system should carry the rest.
- Remove low-value tasks. Let AI handle your inbox. Agents can track forms. Agents manage scheduling and route refills.
- Design clean handoffs. AI → human → EHR, with clear rules.
- Audit everything. No more "pull 20 random" cases. Check safety, accuracy, and speed. Every single interaction is graded and evaluated in real-time. Issues or bugs are flagged immediately and addressed.
“Everybody in healthcare should be working to the top of their license.”
“Turnkey” Is a Myth—Make the EHR Work for You
Every practice is different. EHRs want to be turnkey. Care isn’t. The answer is a standard plan that keeps evolving.
- Use a living project plan. Update it as payer rules and workflows change.
- Use APIs wisely. Add point solutions where the EHR falls short. Avoid a Franken-stack.
- Ship small, weekly tweaks. Don’t wait for giant quarterly releases.
“Our project plan is turnkey—but always evolving.”
— Rebecca
Revenue Starts at Registration
Denials don’t start with a claim. They start at check-in. Fix the front end to lift the back end.
- AI eligibility + plan match. Read IDs, confirm coverage, and tag PPO/HMO correctly.
- Smart scheduling. Match payer rules to the right slot and service codes.
- Automate prior auth and referrals. Let software chase status and fill forms. Escalate exceptions.
This reduces both front-end and back-end denials simultaneously. Fewer re-touches. Faster payments. Less staff frustration.
Culture Eats Rollouts: Upskill, Don’t Replace
Tools are easy. Change is hard. The blocker is often culture.
- Name your champions. Find respected clinicians to test and teach.
- Create space to learn. Give protected time in a training environment.
- Redeploy people. Don’t sideline “Betty in Records.” Train her to audit AI, work denials, and run QA.
“We can elevate her beyond admin tasks. There’s a management piece—audits, training the AI, and monitoring.”
— Rebecca
A Simple 30-Day Day-0 Plan
Week 1: Stabilize
- Turn on AI for simple calls.
- Daily huddles: top three issues, owners, next steps.
- Create a “red phone” path for urgent escalations.
Week 2: Clean the Front End
- Select your AI eligibility and pick a plan when you register.
- Add scheduling rules tied to payer requirements.
- Verify registration and assess the quality of call triage.
Week 3: Move the Paperwork
- Automate prior auth and referrals with clear exception queues.
- Verify refill requests to the correct EHR inbox and add safety checks / alerts to flag issues.
- Publish your first “what we fixed” note to staff.
Week 4: Lock in Gains
- Review denial trends. Fix the top two root causes.
- Remove one extra manual step from a common workflow.
- Share before/after metrics. Celebrate one team's win.
Metrics That Matter (One Page)
Access
- First-ring answer rate
- AI self-serve rate
- Time-to-human on escalation
Operations
- Average handle time (AI + human)
- % tasks auto-routed to EHR
- Clinician minutes on admin per day
Revenue
- Clean claims rate
- Denial rate (front vs. back end)
- Days in A/R
Experience
- Patient satisfaction after calls
- Clinician satisfaction with inbox load
If a metric doesn’t drive action, cut it.
Safety, Audits, and Guardrails
AI works best with clear limits.
- Define green / yellow / red. What AI can do alone, when to ask a human, and what it must never do.
- Audit small, often. Review low-scoring cases each week. Share findings. Fix fast.
- Document handoffs. Who owns what and how to escalate in one click.
What to Say to the Skeptics
- “Try it in a sandbox.” Let them click before launch.
- “We’ll fix three things each week.” Keep promises small and public.
- “Your voice drives the backlog.” Show how feedback ships.
- “We’re measuring the right things.” Share the one-page dashboard.
Real-World Patterns I’m Seeing
- Voice is the best wedge. Great phone service earns trust fast.
- Eligibility accuracy beats heroics. Get plan types right and everything speeds up.
- Short loops win. Weekly changes beat monthly overhauls.
- People matter. Pay well, lighten the load, show progress. Turnover drops.
- APIs are power tools. Use them to finish the job the EHR started, without chaos.
5 Key Takeaways
- Go-live is Day Zero. Budget 30 days for real-world fixes and visible wins.
- Start at the front door. AI on the phone handles simple needs and routes the rest.
- Protect clinician time. Offload admin so people work at the top of their license.
- Fix registration to fix revenue. Eligibility, plan match, and smart scheduling cut denials.
- Culture makes it stick. Champions, training time, and steady audits turn tools into results.
Final Thoughts
If I could bottle the feeling of a great clinic day, it wouldn’t be “we launched a system.” It would be this: patients received help quickly, clinicians had sufficient time to care, and the team went home on schedule. That’s what Health + AI should deliver.
To see the AI tools we built to make this a reality, visit Pretty Good AI and head over to Ignite Healthcare Solutions to learn about the work Rebecca and team do to make the best of Athena Health EHR.
See you next week,
-kevin
Recent Social Posts
Recent Podcasts
Catch up on past emails here.
2810 N Church St #87205, Wilmington, DE 19802
Unsubscribe · Preferences
