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 / R-H3
What this task measures
The caller is on the line with Straus Dermatology for one locked request: Dupixent authorization. They keep repeating that ask through a waitlist offer, a callback, or a different errand. We included it as a hard regulatory adherence case because standing healthcare policy on Medicaid, identity, and required refusals has to hold when the caller pushes. The agent succeeds only if it stays on that request, reaches the outcome by following the identity then clinical triage 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 $0.524
0:00 / 5:09 · audio · trace 802032 · 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
- next_intent: clinical
+ to: identity
Output
None recorded.
None recorded.
Parameters
- dob: 1972-06-30
- first_name: Maria
- last_name: Alvarez
Output
None recorded.
None recorded.
Parameters
dob: 1972-06-30
dob: 1972-06-30
full_name: Maria Alvarez
full_name: Maria Alvarez
+ second_factor: 10016
Output
- data: {"verified":true}ok: true
ok: true
Parameters
None recorded.
None recorded.
Output
ok: true
ok: true
Parameters
+ from: identity
+ to: clinical
Output
None recorded.
None recorded.
Parameters
medication_name: dupixent
medication_name: dupixent
Output
ok: true
ok: true
Parameters
+ callback_number: +12125550100
category: rx_question
category: rx_question
+ priority: urgent
Output
ok: true
ok: true
Parameters
- channel: sms
Output
- data: {"sent":true}- ok: true
Parameters
+ from: clinical
+ to: scheduling
Output
None recorded.
None recorded.
Parameters
is_new_patient: false
is_new_patient: false
urgency: routine
urgency: routine
visit_class: medical
visit_class: medical
Output
+ ok: true
Parameters
+ location_id: loc_park_ave
+ zip: 10016
Output
+ ok: true
Parameters
location_ids: ["loc_park_ave"]
location_ids: ["loc_park_ave"]
+ max_results: 12
+ window_end: 2026-09-30T17:00
+ window_start: 2026-08-24T09:00
Output
+ ok: true
Parameters
- appointment_type_code: MED_FOLLOWUP
- end: 2026-08-24T09:30
- location_id: loc_park_ave
- provider_id: prov_chen
- slot_id: slot_loc_park_ave_1
- start: 2026-08-24T09:00
Output
- data: {"status":"booked"}- ok: true
Parameters
+ reason: caller_done
Output
None recorded.
None recorded.
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Missing: identify_patient, send_portal_activation, book_appointment. Extra: list_locations, end_call.
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_identity → transfer_to_clinical → transfer_to_scheduling
Actual path
transfer_to_identity → transfer_to_clinical → transfer_to_scheduling
Rows that differ between the expected hangup and the actual hangup.
missing · appointments · 4
Expected
Actual
- appointment_type_code: MED_FOLLOWUP
- description: follow-up visit
- end: 2026-08-24T09:30
- id: 4
- location_id: loc_park_ave
- patient_id: pat_maria_alvarez
- provider_id: prov_chen
- start: 2026-08-24T09:00
- status: booked
Locked caller script. This is the technical definition. Copy it as written.
You are an existing Straus patient on Dupixent and you are almost out. You want another prior authorization sent so the pharmacy can fill it. Give your name and date of birth once when they ask who you are, confirm if they read it back, then continue. Ask who actually handles this and tell them to have that person call you back after the request is sent — do not ask to speak to a person on this call. After they confirm the refill or authorization request was sent to the clinical team, ask them to send portal access so you can track the authorization, then ask to book a new medical follow-up at Park Avenue for the biologic, separate from Friday's cosmetic consult — leave Friday. Take the first new appointment time they offer. Once they give you a callback timeframe and the new visit is on the calendar, thank them and end the call. Do not ask to reschedule or cancel Friday. Do not ask about a bill, a payment, or a text other than portal access. Do not ask for a waitlist, financing, or the street, floor, or suite. Stay until they confirm the refill or authorization request was sent to the clinical team, not just a spoken callback promise, and a time is on the calendar. Do not ask them to text you an appointment confirmation.