What can an AI agent collect during patient intake?
Your team gets a complete intake record before the visit. The patient talks to a PolyAI voice agent by phone, ideally while booking. It asks your questions on symptoms, health history and insurance, accepts answers in any order, follows up on incomplete ones and reads back details like a member ID, and once it's connected to your eligibility source it checks coverage during the call. It transfers anything urgent to your team and writes fields into your electronic health record (EHR) where your system allows.
The questions the agent asks on an intake call
Your intake staff ask the same questions on almost every call. Those answers usually land on a paper or static form and get typed again into the electronic health record (EHR), which means duplicate entry, extra administrative work and room for typos in details like member IDs. The agent asks them in a conversation and lets the patient answer in whatever order they think of things.
New patient details
The agent collects what your registration form asks for:
- Full name, including any name change since their last card.
- Date of birth and contact details.
- Address and preferred pharmacy.
- The reason for the visit, in the patient's words.
- How they were referred, if your team tracks it.
Symptoms and health history
Your team gets the patient's symptoms and history in their own words, without a form. The agent works from your intake script and asks your form's questions in conversation, rephrasing and following up when an answer is incomplete. Each answer shapes what it asks next within your script: for example, if a patient mentions a cough but not how long they've had it, it asks that before moving on. Your clinicians review what it records, since the agent makes no clinical judgments.
Insurance details
Insurance is where phone intake usually goes wrong, because member IDs are long and patients read them off a card in a hurry. When a patient says "it's under my husband's plan through his work," the agent asks for the subscriber's name and date of birth along with the member ID, then reads the ID back so the patient can correct it.
The agent can check coverage during the call and flag a problem to your team before the visit. Insurance checks usually run through an eligibility verification vendor, and sometimes through your EHR, depending on how it's configured.
Intake on the booking call
Where you can, we think intake belongs on the scheduling call, because the patient is already on the line and already thinking about the visit. The agent books the appointment, then asks the intake questions for that visit type, so one call does both.
How Diamond Recovery runs intake by phone
Diamond Recovery treats addiction and mental health conditions across the US, from detox and residential care to outpatient programs, and takes about 500 calls a day. Many of its callers are reaching out at a hard moment. The agent greets them calmly, asks the intake questions and collects personal information and insurance credentials in free-flowing conversation, without making anyone repeat themselves. It then passes the record to a specialist team, who set up the treatment plan.
In partnership with Diamond Recovery, we have the agent handling 74% of those calls end to end, so most callers get what they need in a single conversation. Together we also have it answering more than 40 frequently asked customer service questions, and enrollment has been higher since launch.
The agent does not decide whether a patient is clinically right for a program. It gathers the facts, and your clinicians make that decision. If a caller mentions something urgent, the agent transfers the call to a person on your team, following the escalation and handoff pathways your team defines.
Intake that lands where your team works
Intake data has to reach the system your staff open in the morning. Besides Epic, PolyAI connects to ModMed for specialty scheduling and intake and to athenahealth for ambulatory practices, and Diamond Recovery's agent works with Twilio and Salesforce. If your intake runs somewhere else, the agent connects to it through its API.
An agent built in Dialog Studio (formerly Agent Studio) can also run on web chat and two-way SMS, for patients who would rather type than talk.
Patient data is protected under the HIPAA Security Rule, the US rules for safeguarding health information, and a business associate agreement, a contract that sets out how we handle that data, is available for eligible healthcare deployments. Data handling, retention and model use are governed by your agreement and deployment configuration.
Questions intake teams ask
How is this different from a digital intake form?
The agent holds a conversation and lets the patient answer in any order. When an answer needs one, it asks a follow-up, such as the subscriber's name and date of birth when coverage is under a spouse's plan, as described under insurance details. Details that are easy to mishear are also read back, so the patient can correct them on the spot.
Can the agent verify insurance eligibility?
Yes, once it's connected to your eligibility source. Insurance checks usually run through an eligibility verification vendor, and sometimes through your EHR, depending on how it's configured. The agent then checks coverage during the call and flags a problem to your team before the visit.
Does the agent write intake answers into our EHR?
It passes a complete record to your team, and it writes fields into your EHR where your system allows it and your team has approved them.
Do patients need an app, and what about other languages?
No app is needed. Patients call the same number they use today and talk to the agent, which suits those who are not comfortable with forms or apps. For patients who prefer to type, web chat or two-way SMS works instead. The PolyAI platform supports 73 languages.
What if a patient doesn't have their insurance card?
The agent collects everything else, notes what's missing and passes the record on, so your team knows what to chase before the visit.
What happens if a patient mentions something urgent?
The agent does not diagnose or independently assess clinical urgency. It can transfer the call to a person on your team, and your clinicians decide what happens next.
How should we choose voice AI for patient intake calls?
Run a test call with your own intake script. Listen for how the agent copes when a patient answers out of order, whether it reads back the details that matter, and what it does with a question it can't answer.
How much staff time does this save, and does it help with no-shows?
Your team gets time back for complex and sensitive cases because the agent takes over the routine intake conversations, so nobody retypes answers from a form into the EHR, as the Diamond Recovery results described above show. On no-shows, the complete record and the coverage check land before the visit, so your team can chase missing details or coverage problems ahead of the appointment.
How do we get started, and what does it cost?
PolyAI works with your team from the start: analyzing your real calls, reviewing your intake workflow, designing the conversation with you, connecting your telephony and electronic health record system, testing, then monitoring after launch. How long that takes depends on your integrations, how much of your intake goes in on day one, and your testing and sign-off. Pricing depends on your call volumes and setup, and our pricing page has the details.
Hear it on your own calls
Share the script your front desk reads from today, and in a 30-minute demo we'll run it through a live agent while you listen.