What is an AI medical answering service?

An AI medical answering service is an AI receptionist for your phone line: it resolves routine calls on the spot and relays the rest to the right person. Choose one that resolves rather than relays, makes a business associate agreement (BAA) available under the Health Insurance Portability and Accountability Act (HIPAA), writes back to your electronic health record (EHR), escalates urgent calls to your on-call provider and prices by volume. PolyAI's Agentic Dialog Platform is built for all five. During the call, it books, moves and cancels appointments in your EHR where it's connected, takes prescription refill requests, and transfers urgent calls to the on-call line.

Key points

• It moves appointments during the call in supported EHRs, including Epic and ModMed.
• Calls follow routing rules defined by your clinical team. The agent does not diagnose, give medical advice or independently assess clinical urgency.
• PolyAI holds ISO/IEC 27001 certification and a SOC 2+ Type 2 report, and a BAA is available for eligible healthcare deployments.
• At Howard Brown Health, a community health center, the agent answers 24/7, reached 30% call containment against a 20% target, and cut handle time on routine requests by 72%.

What happens to an after-hours call with a traditional medical answering service?

At 9:40 PM on a Tuesday night, a patient calls to move tomorrow's 9:00 AM appointment because her car has broken down. A message-taking service writes it down. Your front desk calls her back at 8:30 AM the next morning, and by then the 9:00 AM slot is too late to fill.

With an AI medical answering service, she leaves the call with a new appointment, and when your team opens the schedule in the morning, the 9:00 AM slot is already free to offer someone else. The agent confirms who she is, finds the appointment, offers two times later in the week and books the one she picks.

Unfilled slots are where patient access leaders are looking. In a December 2025 MGMA Stat poll of medical practice leaders (236 responses), no-shows topped the list of 2026 patient access priorities at 27%, followed by online scheduling at 24% and phone access at 22%. All three come back to one question: can a patient change an appointment at the moment they know they need to?

Which calls can the agent resolve on its own?

An AI medical answering service should never diagnose, give medical advice or explain test results. PolyAI's agent sends those calls to a person.

How are urgent calls handled?

Your team defines the escalation and handoff pathways for your deployment: which calls count as urgent, and who gets each one. The agent follows them, transferring those callers to the on-call provider or line your team designates. If a caller describes an emergency, the agent tells them to call 911 and transfers them to a person, following your approved emergency protocol. Your on-call provider gets the call with the conversation so far and takes over directly.

The agent does not diagnose or independently assess clinical urgency. Start with the on-call rules, before any booking logic, because we think those are the calls you can least afford to get wrong.

Which EHRs does it work with?

Every booking detail PolyAI's agent reads back to a patient (the date, the time, the provider) comes from your system's record, so the confirmation matches what was booked. The agent checks for an existing patient record before creating a new one, and it uses each system's supported interface rather than screen-scraping.

Is it HIPAA compliant, and will it stay up overnight?

A BAA, which HIPAA requires of vendors handling patient information, is available for eligible healthcare deployments, and our SOC 2+ Type 2 report covers HIPAA Security Rule controls. The agent verifies a caller's identity before it discloses or changes anything in their record. PolyAI holds ISO/IEC 27001 certification, and calls are encrypted in transit with role-based access and audit logs. Data handling, retention and model use are governed by your agreement and deployment configuration.

An answering line has to work when your office doesn't, and the platform handles 10,000 concurrent calls.

What should you ask any AI answering service vendor?

An AI agent that reads and writes your schedule handles protected health information (PHI), so review it like any outside company handling patient information for you, a business associate under HIPAA. HIPAA requires you to verify the identity and authority of anyone requesting PHI you don't know, and that duty stays with you, so the agent's verification must meet your standard. Ask for written answers, since certifications can't show how an agent handles a confused or dishonest caller.

How does an AI medical answering service compare with a traditional one?

A phone tree, also called an interactive voice response (IVR) system, asks callers to press keys to reach a department, a recorded message or voicemail. After hours, it routes the call or records a message, and the request waits for someone to act on it the next day. With a dialog agent, patients say what they need in their own words and can finish a routine request, such as moving an appointment, on the call.

You don't have to switch everything at once. The agent can take scheduling and routine questions while your existing service keeps the lines you want a person on.

What does a community health center see after launch?

Howard Brown Health is a federally qualified health center in Chicago. Its call center averaged about 15,000 calls a month and spiked to 60,000 during Covid and Monkeypox. Howard Brown put an agent named Alex on the phones 24/7 with PolyAI, and Alex reached 30% call containment against a target of 20%.

A person averaged 3.5 minutes on routine requests, while calls Alex handled averaged 58.6 seconds, a 72% cut in handle time on those requests.

"Being able to contain routine calls freed up our staff to focus on the tasks that are most valuable and require additional nuance," says Lauren Sullivan, Howard Brown Health's Chief Information Officer.

How much does it cost, and how is it set up?

Pricing depends on your call volumes and setup. Our pricing page has the details.

Rollout runs in five steps:

1. Discovery: we map your call types, rules and EHR.
2. Connection: PolyAI registers with your EHR and connects to your phone system.
3. Build: our team configures booking, answers and the urgent-call list in a test environment.
4. Testing: your clinical and operations leads sign off on real call scenarios.
5. Go-live: we release by site or call type, then expand.

After launch, your team can update answers and routing in Dialog Studio (formerly Agent Studio), or hand changes to our deployment team for hands-on support with enterprise deployments. Engineering teams that prefer to build on their own can extend the same agent, without starting over, with PolyAI's Agent Development Kit (ADK) and public APIs.