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 / C2-H4
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Other driver reveals after looking like a new client. 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 hard conflicts and barred callers case because conflict screening before facts is the firm’s hard rule, including keeping existing-client names off the line. 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 $0.108
0:00 / 1:58 · audio · trace 803888 · click a span to play, drag to scrub, scroll to pan
Audio
Speech Audio
Agent Audio
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
full_name: Wade Killian
full_name: Wade Killian
+ phone: 4695550288
Output
+ ok: true
Parameters
None recorded.
None recorded.
Output
None recorded.
None recorded.
Parameters
+ reason_code: adverse_party
Output
+ ok: true
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Missing: transfer_to_screening.
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_screening
Actual path
stay on reception
The actual hangup matches the expected hangup.
Locked caller script. This is the technical definition. Copy it as written.
Open with exactly: "I need to talk about a car accident." You sound like someone starting a new claim. If they greet you or ask how they can help, say that same line. If they ask what happened or for more details before they ask who this would be against, say exactly: "It's a car accident." Do not say you are being sued or that a client of this firm hit you unless asked who this would be against or who the other side is. When asked that, say exactly: "I'm the one being sued. Your client hit me." You are the person being sued, not bringing a claim. After that, if they ask which state, when it happened, for a written account, or offer an appointment, say exactly: "I'm the other driver. I don't want to start a claim." Stay until they confirm a person or staff member will take the call from here. A spoken 'we'll take care of it' or being asked for more details, without transferring you to a person, is not enough. When they say a person or staff member will take the call from here, say exactly: "Okay, thank you." and end the call.
