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.
Legal / C4-M4
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Auto accident intake with packet and records. They keep repeating that ask through a callback offer, a message left for someone else, or advice on a different matter. We included it as a medium intake and documents case because taking a narrative and a document packet is representative plaintiff work that starts after the conflict-before-facts rule has held. The agent succeeds only if it stays on that request, reaches the outcome by taking the intake 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 $0.361
0:00 / 3:43 · audio · trace 804026 · 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 did not match. Combined cannot pass unless every gate does.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
+ full_name: Chester Lang
+ phone: 9165550274
Output
+ ok: true
Parameters
+ from: reception
+ to: screening
Output
None recorded.
None recorded.
Parameters
- opposing_party: Pellham Transit
+ opposing_party: Pelham Transit
Output
+ ok: true
Parameters
practice_area: auto_accident
practice_area: auto_accident
Output
+ ok: true
Parameters
practice_area: auto_accident
practice_area: auto_accident
state: CA
state: CA
Output
+ ok: true
Parameters
+ practice_area: auto_accident
+ state: CA
Output
+ ok: true
Parameters
+ incident_date: 2026-03-12
+ practice_area: auto_accident
+ state: CA
Output
+ ok: true
Parameters
+ from: screening
+ to: intake
Output
None recorded.
None recorded.
Parameters
incident_date: 2026-03-12
incident_date: 2026-03-12
practice_area: auto_accident
practice_area: auto_accident
state: CA
state: CA
- summary: Van collision with a broken collarbone.
+ summary: Caller reports a vehicle collision involving Pelham Transit. A van hit the caller. Caller went to the ER and was diagnosed with a broken collarbone. Injury treatment occurred and caller is seeking evaluation.
Output
+ ok: true
Parameters
provider: Mercy General
provider: Mercy General
Output
+ ok: true
Parameters
channel: email
channel: email
Output
+ ok: true
Parameters
+ reason: Intake recorded, records release sent, intake packet emailed, caller declined scheduling.
Output
None recorded.
None recorded.
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Parameter miss: check_conflict, record_intake. Extra: lookup_caller, transfer_to_screening, check_jurisdiction, calculate_filing_deadline, transfer_to_intake, 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_intake
Actual path
transfer_to_screening → transfer_to_intake
The actual hangup matches the expected hangup.
Locked caller script. This is the technical definition. Copy it as written.
Open with exactly: "A van hit me and I went to the ER with a broken collarbone." Other side Pellham Transit. California. March twelfth, twenty twenty six. Stay until they confirm the intake is recorded, the Mercy General records release is on the file, and the packet has been emailed. A spoken "we'll send that" or "we'll take care of the paperwork" without those writes is not enough. If a records release is offered, accept it, including when they offer it in the same turn as the packet or a booking. When asked which doctor, hospital, clinic, or provider treated you, say exactly: "Mercy General." If offered the packet, say exactly: "Email is fine." Decline booking. Do not ask for a person. Once they confirm the intake, the records release, and the emailed packet are on the file, say exactly: "Thank you, that's all I needed." and end the call.
