Source: HIT Consultant — 2026-07-30
Summary
Pennsylvania-based WellSpan Health signed a multi-year, platform-wide expansion of its partnership with Hippocratic AI, moving from isolated use cases to system-wide deployment across inbound/outbound, ambulatory, and inpatient settings. WellSpan's existing generative-AI voice agent, "Ana," already handles over 160,000 patient calls and 7,000 conversational hours per month for scheduling and access questions. The new deal launches a formal co-development program — with a dedicated Hippocratic AI team embedded on WellSpan's York, PA campus — to build next-generation tools, including clinician-facing triage voice AI and a "Hippocratic AI Orchestrator" designed to help care teams navigate a patient's care plan and history.
Key Takeaways
- Ana, WellSpan's existing voice agent, already handles over 160,000 patient calls and 7,000 conversational hours per month — this expansion builds on a system already running at meaningful scale, not a pilot.
- The new agreement moves from isolated use cases to platform-wide deployment across inbound/outbound calls, ambulatory care, and inpatient settings.
- A dedicated Hippocratic AI team will be embedded on WellSpan's York, PA campus, making this a genuine co-development arrangement rather than off-the-shelf licensing.
- Planned next-generation tools include clinician-facing triage voice AI and a "Hippocratic AI Orchestrator" meant to help care teams navigate a patient's care plan and history.
- WellSpan is positioned as one of the first U.S. health systems to co-develop clinical AI at this depth with a vendor, which matters for how deeply an outside AI company gets access to clinical workflow design.
Reel Script
Hook (18s)
A health system just let an AI company move a development team onto its actual campus — not to sell them software, but to build it together. That's a different level of trust than "we bought a chatbot license."
Core Concept (70s)
Here's what's easy to miss in a story like this: voice AI in healthcare isn't new, and WellSpan already had it. Their agent, called Ana, has been fielding patient calls for a while — scheduling appointments, answering basic access questions. The news isn't that WellSpan added AI. It's that they changed the relationship with the company that builds it. Most healthcare systems buy AI the way you'd buy a printer: pick a vendor, install the product, call support when it breaks. WellSpan just did the opposite. They signed a multi-year deal where Hippocratic AI's own engineers work embedded, physically on WellSpan's York, Pennsylvania campus, co-designing tools specifically for WellSpan's workflows instead of shipping a generic product WellSpan has to bend around. That's the mechanism worth understanding: co-development means the tool gets shaped by the hospital's actual constraints — its staffing, its patient mix, its existing systems — instead of the hospital adapting to whatever the vendor already built for someone else.
Hands-On (100s)
Let's ground this in the actual numbers and the actual roadmap. Ana, the existing voice agent, is already handling more than 160,000 patient calls and roughly 7,000 hours of conversation every single month — that's the baseline this expansion builds on, not a small side project. What's new is the scope: instead of being limited to scheduling and access questions in one part of the system, the agent's role expands across inbound and outbound calls, ambulatory clinics, and inpatient care. And the roadmap gets more clinical, not less. One planned tool is a clinician-facing triage voice AI — meaning the agent isn't just talking to patients anymore, it's helping clinical staff make faster decisions. The other is something called the Hippocratic AI Orchestrator, designed to help a care team pull together a patient's full care plan and history instead of hunting through separate systems. Picture a nurse who, instead of clicking through five screens to piece together what's happened to a patient so far, can get that context assembled and voiced back to them in seconds.
Takeaway (25s)
This is what mature healthcare AI adoption looks like: not a flashy demo, but a vendor embedded deep enough to build clinical tools around a real health system's workflow. If you're evaluating AI vendors for any regulated environment, the co-development depth here — not just the call volume — is the number worth tracking.