Will patients pick up the phone if an AI is calling? We analyzed over a million real interactions across post-discharge and chronic care programs - here's what the data actually shows.
When we start working with a new health system, two questions almost always come up first from the clinical team: Will my patients actually pick up the phone? And if they do, will they want to talk to an AI agent?
They're fair questions. The technology is new, and for most clinicians this is the first time they've thought about patient engagement with AI voice outreach. The questions deserve answers grounded in real data, not assumptions. We've analyzed more than 1,000,000 patient calls across post-discharge follow-up, chronic disease management, post-procedure check-ins, and administrative workflows. Here's what we've seen:
What the Numbers Actually Show
Across all program types, average reach holds between 85-93% and average engagement between 94-98%. Patients aren't just picking up, they're completing the conversation.
The finding that surprises people most: age makes no meaningful difference. In the heart failure longitudinal program, patients aged 80-89 showed higher reach than those aged 60-69. Engagement holds steady from patients in their 50s through their 90s, including in our longitudinal heart failure program, where elderly patients have been engaging with the AI for years.
So What Encourages Patients to Engage?
It's a phone call. Not an app, not a portal, not a new login. The barrier to engagement is low, patients answer on the phone they already have, in a format they've used their whole lives.
They were told to expect it. A major driver of engagement wasn't age or diagnosis, it was whether the care team told the patient a call was coming before discharge. Engagement consistently increased when care teams set that expectation beforehand.
The AI is transparent about being an AI, and warm enough that it matters. LOLA introduces herself as an AI at the start of every call, and tells patients there's a clinical team behind her so if she can't help with something, a human will. That combination of transparency, empathy, and a clear escalation path is what patients respond to. The transparency itself becomes a trust signal: patients who know exactly what they're talking to engage more, not less.
It resolves things faster than the alternatives. A medication question, a symptom worry, a rescheduling request, a portal login that isn't working. Most of what patients call about can be handled inside a single five-minute interaction, instead of holding on a nurse line, leaving voicemails, or waiting days for a callback. Once patients realize they get an answer now, completion stops being something we need to chase.
Trust and Privacy
Patient trust comes from the conversation itself. Institutional trust comes from what makes the conversation possible, the architecture, the audit trail, the data flow. The two work together: the experience patients respond to is built on top of a system the institution needs to be able to inspect.
The question we hear most often from CMOs in early conversations isn't whether AI voice is HIPAA-compliant. They assume that. The real question is whether the AI is a black box, whether they'll be able to see what it did, what data it touched, and why it made the calls it made.
Nothing about the system is opaque. All calls are recorded and transcribed. Every decision the AI makes is traceable to a specific moment in the conversation and every piece of data accessed is logged, so compliance teams can pull the full audit trail for any interaction at any time.
That changes how AI voice outreach sits inside the rest of the health system's governance.
What This Means for Health Systems
What patients respond to is the call itself, the format, the transparency, the speed. That response holds steady across age groups, program types, and care settings.
The real challenge for AI post-discharge follow-up and chronic care management sits one layer below: a clinically validated protocol, an audit-ready architecture that fits inside existing governance, and a care team that knows what to do with the flags the AI raises.
When those pieces are in place, the answer to the two questions clinical teams ask us at the start of every deployment lands the same way: Patients pick up. They finish the conversation. Clinical teams stop spending hours on outreach that doesn't connect and start spending those hours on the patients flagged by the calls that did.
That's where this stops being about efficiency and starts being about outcomes: earlier intervention, fewer missed deteriorations, and better care for the patients who would otherwise have slipped through.
The data in this post comes from Tucuvi's 2026 Patient Engagement in AI Voice Follow-Up Report, based on real-world deployment data across multiple care settings. [Download the full report here.]
