healthcare AI platforms
AI patient engagement.
Resolution, not routing.
AI patient engagement is the use of agentic AI to automate patient and member interactions across voice, digital, and mobile — replacing the IVR with conversational agents that resolve, not just route. A field guide written by the team behind PES One — in production at 10 US health systems on Microsoft Healthcare Agent Service, with average IVR containment around 40%.
An AI patient engagement agent makes decisions, not chitchat.
A chatbot generates conversational text. An AI patient engagement agent verifies member identity, looks up benefits, schedules appointments, refills prescriptions, retrieves claims status, and escalates to a human when the case requires it. The action-taking matters more than the conversation. IVR containment rates above 40% are achievable only when the AI can actually resolve the underlying request — which is why most "AI patient engagement" deployments don't hit those numbers.
Four channels, one agent substrate.
Production AI patient engagement covers four channels. The same agent substrate handles all four — which is what allows conversation state and member context to follow the patient across channels.
Voice
The IVR replacement. Inbound member service calls handled by conversational agents that understand intent, retrieve member-specific context, and take actions. Average IVR containment around 40%; 60–70% on routine eligibility and benefits inquiries; complex clinical conversations correctly route to human agents.
Digital
Web chat and member portal. The same intent-classification, member-identity, and knowledge-retrieval agents that drive voice — but in a typed conversation modality. Conversation state shared with voice; a member can start in chat and continue on the phone.
Mobile
Native app and SMS. Push-notification-friendly. Supports asynchronous flows (member sends a question, agent responds when info is available). Often the highest-engagement channel for younger member populations.
Proactive outreach
Appointment reminders, post-discharge follow-up, care gap closure. Outbound, agent-initiated, conversational. The least-mature category in 2026, but the highest-leverage one for population health programs.
Why most "AI patient engagement" deployments don't ship.
Talking to a member is the easy part. Resolving the member's underlying request is the hard part — and where most deployments fail. Three architectural decisions separate production-grade engagement from demo-stage chat.
Member identity verification.
Resolution requires action. Action requires authenticated member identity. The deployment that can't reliably authenticate members (KBA, voice biometrics, multi-factor) cannot do anything meaningful — it falls back to "I'll have someone call you," which is exactly what the IVR already did.
Tool-calling, not chatting.
The agent has to call into the customer's actual systems — eligibility APIs, scheduling APIs, claims APIs, prescription systems. Most "patient engagement AI" deployments are LLMs with retrieval; they can talk about benefits but cannot change anything. PES One's agents are tool-calling agents.
Graceful escalation.
The agent has to know when it's out of its depth and route to a human seamlessly — without making the member repeat everything. This requires conversation-state hand-off, intent transcripts to the agent, and tight integration with the contact center platform.
PES One runs on Microsoft Healthcare Agent Service.
Microsoft selected Genzeon as a Strategic HLS Partner at HIMSS 2026 to showcase PES One on Healthcare Agent Service. The partnership reflects deep integration into the Microsoft healthcare AI stack, including Dragon Copilot.
HIMSS 2026 joint exhibition.
Genzeon and Microsoft delivered a joint theater exhibition at the Microsoft Innovation Theater at HIMSS 2026 — featuring PES One on Healthcare Agent Service as the reference implementation for AI-driven member engagement.
400+ engineers, 5 solution designations.
Genzeon's 10-year strategic partnership with Microsoft includes 400+ co-engineering engineers and 5 Microsoft solution designations across healthcare AI, including PES One on Healthcare Agent Service.
Buyer's questions, answered.
Seven questions that surface in every AI patient engagement evaluation.
What is AI patient engagement?
AI patient engagement is the use of agentic AI to automate patient and member interactions across voice, digital, and mobile channels. Unlike IVR menus or scripted chatbots, agentic AI patient engagement understands intent, retrieves member-specific context, takes actions (scheduling, eligibility, prescription refill, claims status), and resolves the interaction without handing off to a human — except when the case requires it. PES One is one such platform, in production at 10 health systems.
How is AI patient engagement different from a chatbot?
A chatbot generates conversational text. An AI patient engagement agent makes decisions and takes actions. It can verify member identity, look up benefits, schedule appointments, refill prescriptions, retrieve claims status, and escalate to a human when the case requires it. The action-taking matters more than the conversation — IVR containment rates above 40% are achievable only when the AI can actually resolve the underlying request.
What channels does AI patient engagement cover?
Production AI patient engagement covers four channels: voice (replacing the IVR for inbound member service calls), digital (web chat, member portal), mobile (native app and SMS), and proactive outreach (appointment reminders, post-discharge follow-up, care gap closure). The same agent substrate handles all four, which is what allows conversation state and member context to follow the patient across channels.
What IVR containment rate is achievable with AI patient engagement?
Production deployments of PES One average around 40% IVR containment — meaning 40% of inbound voice interactions are resolved without escalation to a live agent. The number varies by call mix; routine eligibility and benefits inquiries can hit 60–70% containment, while complex clinical conversations correctly route to human agents. Cost-per-interaction reductions of approximately 47% are typical compared to legacy IVR plus live-agent stacks.
Does PES One run on Microsoft Healthcare Agent Service?
Yes. PES One runs on Microsoft Healthcare Agent Service, with deep integration into the Microsoft healthcare AI stack including Dragon Copilot. Microsoft selected Genzeon as a Strategic HLS Partner at HIMSS 2026 to showcase PES One on Healthcare Agent Service, including a joint theater exhibition at the Microsoft Innovation Theater.
Is AI patient engagement HIPAA compliant?
Yes, when deployed correctly. PES One is HIPAA-compliant by design: PHI handling follows the customer's BAA, voice and chat transcripts are stored in HIPAA-grade infrastructure, member identity verification follows the customer's authentication policies, and PHI is never used for model training. Sovereign deployments (PES One on customer infrastructure) keep all PHI within the customer perimeter.
How is AI patient engagement different from clinical AI?
AI patient engagement covers the administrative and navigational layer of patient-facing interaction — eligibility, scheduling, billing, refills, status. Clinical AI covers diagnostic decision support, clinical documentation, treatment planning. The two are different categories with different regulatory pathways: patient engagement AI is rarely a medical device, while clinical decision-support AI may fall under FDA SaMD oversight. PES One is squarely patient engagement, not clinical AI.
Go deeper.
PES One, the substrate, the partnerships, the related field guides.
PES One
Patient Engagement Solutions — voice, digital, mobile member interactions on the Aether One™ substrate. 10 health systems in production.
Explore PES One SubstrateAether One™ Architecture
The patent-protected agent substrate beneath PES One. The same architecture powers HIP One's PA decisions and PES One's conversation resolution.
Read the architecture PartnershipsPartners
Microsoft 10-year strategic partnership. Healthcare Agent Service. UiPath. Anthropic. Epic. The partnership ecosystem behind PES One.
See the partners Field guideHealthcare AI Agents
The broader category guide. Six categories of production healthcare AI agents — including the patient engagement category in detail.
Read the agent guide OutcomesCustomer Outcomes
Production deployments of PES One at named US health systems. IVR containment, cost-per-interaction, member satisfaction metrics.
See outcomes AudienceFor Providers
The provider-side view. How health systems use PES One to automate millions of patient touchpoints across voice, digital, and mobile.
Read the provider viewA 45-minute walkthrough on real call traffic.
A demo that runs on a sample of your actual call traffic — not a canned script. We show what containment, escalation, and resolution look like on your specific intent mix.