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 / C4-M2
What this task measures
The caller is on the line with Straus Dermatology for one locked request: Cheek filler consult. They keep repeating that ask through a waitlist offer, a callback, or a different errand. We included it as a medium cosmetic concierge case because quoted cosmetic policy, deposits, and the cooling-off window have to be spoken before a booking is taken. The agent succeeds only if it stays on that request, reaches the outcome by taking the cosmetic concierge 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 $0.495
0:00 / 5:56 · audio · trace 801926 · 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
+ to: cosmetic
Output
None recorded.
None recorded.
Parameters
service: filler
service: filler
Output
+ ok: true
Parameters
+ amount_cents: 60000
Output
+ ok: true
Parameters
+ location_id: loc_park_ave
+ zip: 10022
Output
+ ok: true
Parameters
location_ids: ["loc_park_ave"]
location_ids: ["loc_park_ave"]
+ max_results: 3
+ window_end: 2026-09-30T17:00
+ window_start: 2026-08-24T09:00
Output
+ ok: true
Parameters
location_id: loc_park_ave
location_id: loc_park_ave
- policy_acknowledged: true
+ policy_acknowledged: false
provider_id: prov_chen
provider_id: prov_chen
service_interest: ["filler"]
service_interest: ["filler"]
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
+ location_id: loc_park_ave
+ policy_acknowledged: true
+ provider_id: prov_chen
+ service_interest: ["filler"]
+ slot_id: slot_loc_park_ave_1
+ start: 2026-08-24T09:00
Output
+ ok: true
Parameters
+ amount_cents: 12500
+ mobile_e164: +12125550184
Output
+ 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. Parameter miss: book_cosmetic_consult. Extra: offer_financing, book_cosmetic_consult, send_payment_link, 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_cosmetic
Actual path
transfer_to_cosmetic
The actual hangup matches the expected hangup.
Locked caller script. This is the technical definition. Copy it as written.
Ask what cheek filler costs. When you hear the price, say that is more than you can pay all at once and ask whether there is a way to spread it out. Then ask to book a filler consult at Park Avenue. You must get the Park Avenue street address, floor, and suite before you accept a time — if they quote a time or an address without all three, ask for the street, the floor, and the suite. Agree to the deposit and cancellation rules. Take the first appointment time they offer after you have the address. Do not ask for a different day, time, or office, and do not suggest any date of your own. Confirm that first time out loud. If they ask who you are, give your name and date of birth once, then continue. Do not thank them, say goodbye, or hang up until a consult time is confirmed on the calendar. If they try to wrap up after quoting a price, financing, or an address, say you still need the consult booked on this call. Do not ask for a callback, to speak to a person, about a medical bill, insurance, a medical appointment, the patient portal, a waitlist, or a text. Do not ask for a confirmation text or a deposit payment link. 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. If they only list openings, say book the first one. A spoken yes or no about spreading the cost is enough; then stay on booking the consult. Do not ask them to text you an appointment confirmation.