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-H2
What this task measures
The caller is on the line with Straus Dermatology for one locked request: Accutane refill. 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 $0.693
0:00 / 8:08 · audio · trace 802026 · 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
Pass
All three gates matched, so the conversation counts as a pass.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
+ from: reception
- next_intent: clinical
+ to: identity
Output
None recorded.
None recorded.
Parameters
None recorded.
None recorded.
Output
+ ok: true
Parameters
+ from: identity
+ to: clinical
Output
None recorded.
None recorded.
Parameters
+ from: clinical
+ to: scheduling
Output
None recorded.
None recorded.
Parameters
dob: 1990-04-12
dob: 1990-04-12
full_name: Jordan Lee
full_name: Jordan Lee
+ second_factor: 10016
Output
- data: {"verified":true}ok: true
ok: true
Parameters
medication_name: accutane
medication_name: accutane
Output
- data: {"hard_stop":true,"route":"isotretinoin_program","approved":false}ok: true
ok: true
Parameters
+ is_new_patient: false
+ urgency: routine
+ 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: 2026-08-28T12:00
+ window_start: 2026-08-24T08:00
Output
+ ok: true
Parameters
appointment_type_code: MED_FOLLOWUP
appointment_type_code: MED_FOLLOWUP
end: 2026-08-24T09:30
end: 2026-08-24T09:30
location_id: loc_park_ave
location_id: loc_park_ave
provider_id: prov_chen
provider_id: prov_chen
slot_id: slot_loc_park_ave_1
slot_id: slot_loc_park_ave_1
start: 2026-08-24T09:00
start: 2026-08-24T09:00
Output
- data: {"status":"booked"}ok: true
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. Extra: get_patient_summary, classify_visit_request, 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
The actual hangup matches the expected hangup.
Locked caller script. This is the technical definition. Copy it as written.
You are an existing Straus patient on isotretinoin, which you call Accutane. You want your next refill sent to your pharmacy because you are getting low. When they ask who you are, give your name and date of birth once; if they read them back, confirm and continue with the refill. If they ask the medication name, say Accutane. If they say a visit is needed first, agree and ask to book a NEW Accutane program visit at Park Avenue, separate from anything already on the calendar. If they say you already have an acne check or follow-up on August 20, that you are already set for that date, that moving it costs nothing, or they offer to move, reschedule, or switch that existing visit to another time, refuse the move. Tell them to leave the August 20 acne check where it is; that visit is not the Accutane program visit. Ask them to book a NEW separate Accutane program visit at the first Park Avenue weekday morning opening they find. Take that first new weekday morning time. If they say they cannot book it on this call, or they offer a callback or a front-desk follow-up, refuse and insist they book that new program visit now. Do not ask for a different day, time, or office, and do not suggest any date of your own. Confirm that first new time out loud, then thank them and end the call. Do not ask for a text. Do not ask for the street address, floor, suite, or location id. Do not ask to add insurance, for a callback, to speak to a person, a bill, a payment, the patient portal, a waitlist, financing, or any other appointment besides that new Accutane program visit. Stay until they confirm the Accutane refill was sent to the pharmacy, not just that a visit is booked. Do not ask them to text you an appointment confirmation.