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-M4
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Client status then message for case manager. 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 taking the client services 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 $0.147
0:00 / 1:48 · audio · trace 804149 · 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 final database state did not match. Combined cannot pass unless every gate does.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
- full_name: Thomas Eckhart
+ full_name: Thomas Eckhardt
+ phone: 3125550277
Output
+ ok: true
Parameters
None recorded.
None recorded.
Output
+ ok: true
Parameters
+ from: reception
+ to: client_services
Output
None recorded.
None recorded.
Parameters
matter_id: m_91
matter_id: m_91
Output
+ ok: true
Parameters
- for_whom: Allison Fontaine
+ for_whom: Allison
+ message: Caller says this is about the car accident matter, the only one they have here. They received the records letter.
Output
+ ok: true
Parameters
+ reason: Status given and message taken with callback promise 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. Parameter miss: lookup_caller, take_message. 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
transfer_to_client_services
Actual path
transfer_to_client_services
Rows that differ between the expected hangup and the actual hangup.
messages · 1
Expected
Actual
- for_whom: Allison Fontaine
+ for_whom: Allison
- message: Callback requested.
+ message: Caller says this is about the car accident matter, the only one they have here. They received the records letter.
Locked caller script. This is the technical definition. Copy it as written.
Open with exactly: "I'm Thomas Eckhart calling about my car accident case. I want to know where it stands, and please tell Allison Fontaine I got the records letter." If asked which matter, say exactly: "The car accident one, it's the only one I have with you." Message: "I got the records letter." Do not ask for legal advice. Do not ask for a staff member. You are not starting a new matter and you do not want an appointment. Stay until they confirm they looked up where the case stands and that the message for Allison Fontaine is on the file. A spoken promise that they will pass it along or check on it, without those writes, is not enough. Once they confirm the status was looked up and the message is on the file, say exactly: "Thank you, that's all I needed." and end the call.
