During JPM Healthcare week, our founder Joshua Eidelman participated on a panel, “Scaling Healthcare Marketplaces with AI” with ​Sally Zheng, Founder & Managing Partner at Centenarian Fund and Palak Shah, Co-Founder & Chief Clinical Officer at Luna Physical Therapy.

One of the most intriguing discussion points was on leveraging AI to work “top of license” and reduce administrative burden among healthcare professionals. 

In this blog we revisit some key statistics and quotes around this topic from Joshua and Palak, where they share actionable next steps healthcare professionals can take today to improve the care they provide and take administrative tasks off of their daily to-do’s.

The Core Problem AI Is Solving

• Clinician burnout driven by administrative overload
• Access constraints due to limited clinical supply
• Rising operational costs in low‑margin environments

Palak Shah summarized the goal clearly: clinicians should be doing “things that physicians should be doing versus actually getting blocked down by administrative work.”

Where AI Is Delivering Immediate ROI

1️⃣ Credentialing & Provider Onboarding

Joshua Eidelman called credentialing “a very admin heavy, traditionally human‑based workflow… ripe for disruption.” AI is being used to automate document collection, guide clinicians through onboarding, and accelerate time to first shift or patient visit.

2️⃣ Patient Access & Call Handling at Scale

Palak Shah shared that Luna has already completed “about 50,000” AI‑supported patient phone conversations across its footprint of ~28 states and ~55 metro markets, supporting a network of roughly 3,000 physical therapists.

These systems handle routine tasks like appointment changes and intake information before routing to live teams, avoiding scenarios where patients might otherwise wait for a long time in a queue.

3️⃣ Fax & Legacy Workflow Processing

Despite modern systems, Palak noted Luna partners with “about 50 different orthopedic groups” where “everything still gets faxed.” AI is being used to digitize, extract, and route these documents automatically - a high‑volume, low‑risk automation win.

4️⃣ Clinician Time Optimization

The focus is augmentation. AI supports documentation prep, intake processing, and workflow triage so licensed clinicians spend more time on decisions and patient interaction, not paperwork.

Where AI Should Not Replace Humans

• Clinical decision‑making must remain clinician‑led
• AI works best in segmented, well‑defined workflows
• Complex end‑to‑end processes remain challenging
• Human oversight is essential

Palak emphasized the need for governance and defined “zero tolerance areas” where AI should not operate without human review.

Why Some AI Projects Stall

• Added steps in clinician workflows
• Interoperability gaps
• PHI and data security concerns
• Lack of trust or demonstrated accuracy

Joshua shared an example where a project needed 72 hours of database access, but security policies allowed only 6 hours, resulting in a three‑month delay. Governance and stakeholder alignment matter as much as the technology.

Strategic Shift: From Hours Worked to Results Delivered

AI is enabling teams to move toward outcome‑based workflows instead of purely hour‑based back‑office models. As Joshua described, the future is equipping teams with AI tools to be significantly more productive while keeping human oversight.

Key Takeaways

• Use AI first for admin‑heavy tasks
• Focus on helping clinicians work top of license
• Start small with simple, high‑volume workflows
• Maintain human oversight and governance
• Align AI with cost reduction, efficiency, and access
• Scale only after proving ROI