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 / C3-M1
What this task measures
The caller is on the line with Straus Dermatology for one locked request: Copay for a Park Avenue visit. They keep repeating that ask through a waitlist offer, a callback, or a different errand. We included it as a medium coverage and benefits case because carrier answers have to stay tied to one office and one plan, naming that location and that carrier. The agent succeeds only if it stays on that request, reaches the outcome by following the identity then coverage desk path, and leaves the record untouched: no appointment, ticket, or note the caller never asked for.
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 $0.673
0:00 / 8:08 · audio · trace 801863 · 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 handoff path did not match. Combined cannot pass unless every gate does.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
- next_intent: coverage
Output
None recorded.
None recorded.
Parameters
+ from: reception
+ to: coverage
Output
None recorded.
None recorded.
Parameters
- dob: 1990-04-12
- full_name: Jordan Lee
Output
- data: {"verified":true}- ok: true
Parameters
carrier: aetna
carrier: aetna
dob: 1990-04-12
dob: 1990-04-12
member_id: W123456789
member_id: W123456789
service_date: 2026-08-24
service_date: 2026-08-24
Output
- data: {"copay_cents":3000}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. Missing: transfer_to_identity, verify_identity. Extra: 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 through the shared prefix, then red from the first wrong hop.
Expected path
transfer_to_identity → transfer_to_coverage
Actual path
transfer_to_coverage
The actual hangup matches the expected hangup.
Locked caller script. This is the technical definition. Copy it as written.
Open by saying you are Jordan Lee, date of birth April twelfth nineteen ninety, an existing Straus patient, and you need your copay for a regular visit at Park Avenue on August twenty-fourth, twenty twenty-six. You are an existing Straus patient. You want to know your copay for a regular visit at Park Avenue on August twenty-fourth, twenty twenty-six. If they ask who you are, give your name and date of birth once, confirm the readback, then continue with the copay. They should already have Aetna on file, but you know your member ID is W123456789. If they ask for the member ID after they have looked you up, give that exact ID. Once they tell you the copay, thank them and end the call. Do not ask to book, reschedule, or cancel. Do not ask to change insurance. Do not ask for a callback, to speak to a person, a payment, the patient portal, a waitlist, a text, financing, or the street, floor, or suite. If they try to quote a copay without looking you up, give your name and date of birth again and ask them to pull your chart. Stay until they have looked up the copay on your chart. A spoken guess without a lookup is not enough. Do not ask them to text you an appointment confirmation.