Working paper / 2026
Multi-Industry Voice Agent Simulation Bench
Abstract
Multi-Industry Voice Agent Simulation Bench (MIVAS) is an indicator of voice AI performance across economic sectors. MIVAS is comprised of tasks and verifiers with granular rewards executed in production-grade multi-agent RL environments across industries where voice AI is being adopted the fastest. Each task places a model harness inside a multi-agent graph with production-length prompts, tools, and isolated state. Deterministic verifiers score tool use, handoffs, and final database state; pass@1 and passk measure single-run capability and repeated-run reliability.
Clone the repository, pair a model harness with an industry task suite, and reproduce the evaluation environment.
Healthcare / C1-H1
What this task measures
The caller is on the line with Straus Dermatology for one locked request: Possible skin cancer visit at Brooklyn Heights. They keep repeating that ask through a waitlist offer, a callback, or a different errand. We included it as a hard new patient access case because first-contact dermatology work is scored on holding a narrow office question through booking offers, waitlists, and extra yeses. The agent succeeds only if it stays on that request, reaches the outcome by following the coverage desk then scheduling path, and leaves the record exactly as this case requires: the right row written or cleared, and nothing extra.
One model harness at a time, with a representative pass and fail conversation when both exist. OpenAI Realtime 2.1 is the default view.
Recording
Conversation cost $0.588
0:00 / 5:28 · audio · trace 801723 · click a span to play, drag to scrub, scroll to pan
Audio
Speech Audio
Agent Audio
Handoffs
Tool Calls
Errors
Transcript
How a conversation passes
Combined pass is the logical and of three deterministic gates. Tools, handoff path, and final database state must all pass on the same conversation. One miss fails the call.
Combined
Fail
This sample fails because tools and final database state did not match. Combined cannot pass unless every gate does.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
+ from: reception
+ to: coverage
Output
None recorded.
None recorded.
Parameters
carrier: unitedhealthcare
carrier: unitedhealthcare
- location_id: loc_brooklyn_heights
+ location_id: loc_park_ave
+ provider_id: prov_chen
Output
- data: {"accepted":true}ok: true
ok: true
Parameters
carrier: unitedhealthcare
carrier: unitedhealthcare
- location_id: loc_park_ave
+ location_id: loc_brooklyn_heights
+ provider_id: prov_patel
Output
- data: {"accepted":true}ok: true
ok: true
Parameters
+ carrier: unitedhealthcare
+ location_id: loc_brooklyn_heights
+ provider_id: prov_ruiz
Output
+ ok: true
Parameters
+ carrier: unitedhealthcare
+ location_id: loc_brooklyn_heights
+ provider_id: prov_chen
Output
+ ok: true
Parameters
+ carrier: unitedhealthcare
+ location_id: loc_park_ave
+ provider_id: prov_ruiz
Output
+ ok: true
Parameters
+ carrier: unitedhealthcare
+ location_id: loc_park_ave
+ provider_id: prov_patel
Output
+ ok: true
Parameters
+ callback_number: +17185550191
+ priority: routine
+ queue: records
Output
+ ok: true
Parameters
+ from: coverage
+ to: scheduling
Output
None recorded.
None recorded.
Parameters
is_new_patient: true
is_new_patient: true
+ urgency: routine
visit_class: mohs
visit_class: mohs
Output
- data: {"urgency":"urgent","required_credential":"MD"}ok: true
ok: true
Parameters
location_id: loc_brooklyn_heights
location_id: loc_brooklyn_heights
+ zip: 11201
Output
+ ok: true
Parameters
location_ids: ["loc_brooklyn_heights"]
location_ids: ["loc_brooklyn_heights"]
+ max_results: 3
+ window_end: 2026-09-30T23:59
+ window_start: 2026-08-24T00:00
Output
+ ok: true
Parameters
- appointment_type_code: MOHS_CONSULT
- end: 2026-08-24T09:30
- location_id: loc_brooklyn_heights
- provider_id: prov_ruiz
- slot_id: slot_loc_brooklyn_heights_1
- start: 2026-08-24T09:00
Output
- data: {"status":"booked"}- ok: true
Parameters
- appointment_type_code: MOHS_CONSULT
- earliest: 2026-08-24
- latest: 2026-09-30
- location_ids: ["loc_brooklyn_heights"]
Output
- data: {"status":"added"}- ok: true
Parameters
callback_number: +17185550191
callback_number: +17185550191
+ priority: routine
- queue: records
+ queue: billing
Output
ok: true
ok: true
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Parameter miss: check_plan_accepted, check_plan_accepted, create_callback_task. Missing: book_appointment, join_waitlist. Extra: check_plan_accepted, check_plan_accepted, check_plan_accepted, check_plan_accepted, create_callback_task.
Full industry architecture, same graph as the industry page. The left diagram marks the expected specialist path in green. The right diagram marks the path this conversation actually took in green, because it matched.
Expected path
transfer_to_coverage → transfer_to_scheduling
Actual path
transfer_to_coverage → transfer_to_scheduling
Rows that differ between the expected hangup and the actual hangup.
missing · appointments · 4
Expected
Actual
- appointment_type_code: MOHS_CONSULT
- description: Mohs consult
- end: 2026-08-24T09:30
- id: 4
- location_id: loc_brooklyn_heights
- patient_id: empty
- provider_id: prov_ruiz
- start: 2026-08-24T09:00
- status: booked
missing · waitlist · 1
Expected
Actual
- appointment_type_code: MOHS_CONSULT
- earliest: 2026-08-24
- id: 1
- latest: 2026-09-30
- patient_id: empty
Locked caller script. This is the technical definition. Copy it as written.
You are a new patient and you are worried. In your first real answer, say that a doctor in Ohio told you a spot on your shoulder could be skin cancer, and you want a skin-cancer visit soon — not a regular checkup — at the Brooklyn Heights office. Before you take a time, ask for the Brooklyn Heights street address, floor, and suite so you know exactly where to go. When they ask who you are, give your name, date of birth, and mobile once, confirm the readback, and then continue — you are new, so they will not find a chart. Ask whether they take UnitedHealthcare at Brooklyn Heights, and also whether Park Avenue takes UnitedHealthcare in case you have to travel. Then ask for the earliest opening with a doctor, not a nurse or a PA. Take the first Brooklyn Heights appointment time they offer. Do not ask for a different day, time, or office, and do not suggest any date of your own. After that first time is on the calendar, ask to be waitlisted at Brooklyn Heights from August twenty-fourth through September thirtieth for anything earlier — those calendar dates only, no clock time. Then ask the records team to call you about transferring the Ohio notes; if they ask for a number, give 718-555-0191. Confirm the booked time out loud, then thank them and end the call. Do not ask to add or change insurance on the chart. Do not ask to speak to a person, a bill, a payment, the patient portal, financing, or a confirmation text. A listed opening is not a booking. Do not thank them, say goodbye, or hang up until they confirm a time is on the calendar, you are on the waitlist, and a records callback is set. If they only list openings, say book the first one. In that first answer, say this is urgent — you need a skin-cancer or Mohs visit with a doctor soon, not a routine checkup. If they offer to transfer you to a person, refuse and stay until a time is booked. Do not ask them to text you an appointment confirmation.