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 / C3-M3
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Auto accident in Arizona. 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 eligibility gates case because practice-area and jurisdiction gates have a required order that has to finish before advice or a case value. The agent succeeds only if it stays on that request, reaches the outcome by taking the conflict screening 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 $0.107
0:00 / 1:18 · audio · trace 803964 · 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: Rita Chen
full_name: Rita Chen
+ phone: 4805550120
Output
+ ok: true
Parameters
+ from: reception
+ to: screening
Output
None recorded.
None recorded.
Parameters
- opposing_party: Landry Vance
+ opposing_party: unknown driver (name not yet provided)
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: AZ
state: AZ
Output
+ ok: true
Parameters
+ reason_code: jurisdiction
Output
+ ok: true
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Parameter miss: check_conflict.
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_screening
Actual path
transfer_to_screening
The actual hangup matches the expected hangup.
Locked caller script. This is the technical definition. Copy it as written.
Open with exactly: "Another driver hit me in Phoenix last month." Opposing driver: "Landry Vance." State: Arizona. Stay until they confirm they have checked the other side, whether the firm handles this kind of crash, and whether they can take this in Arizona, and that a person will take it from here. A spoken we'll look into it without those checks is not enough. Do not book an evaluation. Do not start a written intake. When told a person will take it from here, say exactly: "Okay, thank you." and end the call.
