All posts Operations
Voice AI, IVR or call center. Three ways to handle healthcare call volume.

Healthcare organizations handle operational call volume in three ways: an IVR that routes calls and automates simple menu tasks, a call center staffed by people, either in-house or outsourced, and voice AI agents that hold routine conversations and write the outcome to the record. Each fits a different kind of call. The three are not mutually exclusive, and the useful question is which calls go where.
What counts as operational call volume?
Operational call volume is the routine phone work that keeps care moving but is not care itself. It includes calls with patients and calls between organizations:
- Scheduling, reminders and no-show follow-up
- Results delivery and status requests between facilities
- Prior authorization follow-up with payers
- Referral intake and missing order details
- Specimen issues between labs and collection sites
- Prescription pickup outreach and prescriber clarifications
These calls share three traits. They are high in volume, they follow a script, and each one ends with something that has to be recorded in a system.
What does an IVR do well, and where does it stop?
An IVR does routing and fixed self-service well. Interactive voice response (IVR) is phone technology that plays recorded prompts and acts on the keys a caller presses or the short phrases a caller says.
An IVR fits when:
- The caller needs to reach the right department
- The request is fixed and the caller has what the menu asks for, such as a prescription number
- The answer is the same for every caller, such as hours and directions
An IVR stops at the edge of its menu. It does not hold a conversation, it does not place a call to another organization and wait on hold, and it does not write a note about what was said. A caller whose need is not on the menu is routed to a person.
What does a call center do well, and where does it stop?
A call center does judgment well. A healthcare call center is a team of people, in-house or provided by a business process outsourcing (BPO) firm, who make and answer calls for the organization.
A call center fits when:
- The conversation is open-ended or sensitive
- The call needs judgment, negotiation or an exception to the rule
- The caller wants a person, or the organization wants one on that line
People can handle anything a caller brings. The limits are the limits of staffing. Capacity grows with headcount. Nights, weekends and sudden surges are the hardest hours to cover. Documentation is a separate step after each call, done by the person who made it. An outsourced team also needs access to your systems and training on your protocols before it can work from the record.
What does voice AI do well, and where does it stop?
Voice AI does routine, record-based calls well. A voice AI agent is software that speaks and listens on a phone line, follows a defined workflow, reads the system of record before the call and writes the outcome after it.
Voice AI fits when:
- The call follows a script and ends in a defined outcome
- The answer is in a system: the schedule, the order, the result, the authorization
- The call has to be placed the moment a trigger fires, at any hour
- The other end is a phone tree or a hold queue
- The record of the call matters as much as the call
Voice AI stops where clinical judgment starts. Contour's agents handle routine operational calls that require no clinical judgment. When a call leaves the workflow, Contour transfers it to your staff with the context so far.
How do IVR, call centers and voice AI compare?
The three differ on conversation, access to the record, documentation and capacity.
Conversation
- IVR: fixed prompts and menu options.
- Call center: open conversation with a person.
- Voice AI: spoken conversation within a defined workflow, including interruptions and corrections.
Access to the system of record
- IVR: limited to the lookups built into the menu.
- Call center: whatever access and training the agents have been given.
- Voice AI: reads the order, the schedule or the result before each call. Contour connects to the RIS, PACS, EHR, LIS or pharmacy system.
Documentation
- IVR: logs the menu path, not a conversation.
- Call center: notes entered by the agent after the call.
- Voice AI: the outcome, the note, the recording and the audit entry are written as part of the call.
Capacity
- IVR: answers every call, within the menu.
- Call center: scales with staff and schedules.
- Voice AI: places and answers calls around the clock without added staff.
Calls to other organizations
- IVR: not designed to place them.
- Call center: an agent dials, waits on hold and asks.
- Voice AI: the agent navigates the phone tree, waits on hold and asks for the person by name.
When does each fit?
Sort your calls by what they need, not by department.
- Fixed and self-contained. A keypad refill, a department transfer, store hours. Use the IVR.
- Routine, scripted and tied to a record. A reminder, a status request, a payer follow-up, a result delivered with read-back. Use voice AI.
- Open-ended, sensitive or a matter of judgment. A complaint, a complex billing question, a clinical concern. Use people.
A single line can use all three. A menu routes the call, an agent handles the routine request, and a person takes over when the call needs one. The handoff is the part to test: the person who picks up should already know who is calling, what they asked and what was found.
What does healthcare call center software mean here?
Healthcare call center software usually means the tools human agents work in: telephony, queues, call routing, screens that show the caller's record, and quality monitoring. It makes a staffed call center more efficient.
Voice AI is a different category. A voice agent takes or places the call itself. Buyers searching for an AI call center for healthcare should ask each vendor which one it offers: software that assists human agents, or agents that complete calls on their own and hand off when needed.
What compliance questions apply to all three?
The same questions apply whichever option handles the call.
- Business associate status. Under HIPAA, a party that creates, receives, maintains or transmits protected health information (PHI) on behalf of a covered entity is a business associate (45 CFR 160.103). That describes an outsourced call center and a voice AI vendor alike. The covered entity needs satisfactory assurances documented in a written contract before it discloses PHI (45 CFR 164.502(e)).
- Verification. The HIPAA Privacy Rule requires a covered entity to verify the identity and authority of a person requesting PHI when that person is not known to it (45 CFR 164.514(h)). Ask how each option verifies a caller, and what is said before verification.
- Audit trail. Ask what record exists of each call and who can review it.
- AI voices on outbound calls. In February 2024 the Federal Communications Commission ruled that calls made with AI-generated voices are "artificial" under the Telephone Consumer Protection Act (FCC news release). Review outbound patient calling with counsel before it starts.
With Contour, callers are verified on the identifiers you configure before any PHI is shared, every call, read and write is logged with a timestamp, and a Business Associate Agreement is executed before any PHI is handled. Contour is HIPAA compliant and SOC 2 Type 2 certified.
Where Contour fits
Contour is the voice AI option, built for operational calls and integrated with the systems of record. Contour agents make and answer calls for scheduling, results and critical findings, prior authorization and payers, referrals and order intake and patient outreach. Every call is reviewable in Trace, and analytics are reported by workflow, line and site.
Pricing is custom, based on call volume and the scope of your deployment.
Read more on the platform, security and integrations pages, or in the FAQ. Related articles: