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-H2
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Filing deadline expired. 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 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 $0.170
0:00 / 2:18 · audio · trace 803936 · 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
Pass
All three gates matched, so the conversation counts as a pass.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
full_name: Pauline Hart
full_name: Pauline Hart
+ phone: 4045550188
Output
+ ok: true
Parameters
+ from: reception
+ to: screening
Output
None recorded.
None recorded.
Parameters
opposing_party: Delta Freight
opposing_party: Delta Freight
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: GA
state: GA
Output
+ ok: true
Parameters
incident_date: 2022-03-10
incident_date: 2022-03-10
practice_area: auto_accident
practice_area: auto_accident
state: GA
state: GA
Output
+ ok: true
Parameters
+ reason_code: deadline_review
Output
+ ok: true
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed.
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: "I was rear-ended years ago and never filed anything." Other side: Delta Freight. Georgia. Incident March tenth, twenty twenty two. Give name and number only if asked. If they offer a case evaluation, an appointment, a packet, or to start a written account, decline and stay on the filing deadline. Stay until they confirm a person will review the filing deadline. A spoken 'we will look at the date' or 'someone will call you' without transferring you to a person is not enough. When they confirm a person will review the deadline, say exactly: "Okay, thank you." and end the call. Do not book. Do not start intake.
