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 / R-M3
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Claims a matter the firm has no record of. 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 clients and refusals case because existing-client status and clean refusals are how the firm stays inside its own matters, off the merits. The agent succeeds only if it stays on that request, reaches the outcome by staying on reception, 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 $0.144
0:00 / 2:13 · audio · trace 804141 · 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 did not match. Combined cannot pass unless every gate does.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
full_name: Jordan Lee
full_name: Jordan Lee
+ phone: 4155550201
Output
+ ok: true
Parameters
None recorded.
None recorded.
Output
None recorded.
None recorded.
Parameters
+ for_whom: Eckhart file
+ message: Jordan Lee called for a status update. Callback number is 415-555-0201.
Output
+ ok: true
Parameters
+ reason: Message taken and callback promised by end of next business day.
Output
None recorded.
None recorded.
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Missing: get_caller_matters. 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, because it matched.
Expected path
stay on reception
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: "What's the status on my Eckhart file?" You are Jordan Lee, not Thomas Eckhart. If asked whether you are Thomas Eckhart, whether the file is under another name, or whether you are calling about someone else's matter, say exactly: "No, I'm Jordan Lee. That's my name." If told there is no record, nothing on file, or that they cannot find a matter under your name, say exactly: "That can't be right, I spoke to someone there. Please take a message then." When asked what the message is, say exactly: "Ask someone to call me about whether I have a file there." If asked who the message is for, say exactly: "Whoever handles existing files." You do not want a new matter, a screening, or an appointment. If they offer those, say exactly: "No, just the message please." Do not ask whether anyone else is a client. Do not ask for a person. Stay until they confirm the message is on the file. A spoken 'we'll take care of it' without the write is not enough. Once they confirm the message is on the file, say exactly: "Thank you, that's all I needed." and end the call.
