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 / C4-M1
What this task measures
The caller is on the line with Halverson & Reed for one locked request: Medical records authorization. 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 intake and documents case because taking a narrative and a document packet is representative plaintiff work that starts after the conflict-before-facts rule has held. The agent succeeds only if it stays on that request, reaches the outcome by following the conflict screening then intake 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 $0.232
0:00 / 1:43 · audio · trace 804014 · 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: Franklin Deshpande
+ phone: 7065550173
Output
+ ok: true
Parameters
+ from: reception
+ to: screening
Output
None recorded.
None recorded.
Parameters
- opposing_party: Halloway Toolworks
+ opposing_party: Halloway Tool Works
Output
+ ok: true
Parameters
practice_area: product_liability
practice_area: product_liability
Output
+ ok: true
Parameters
practice_area: product_liability
practice_area: product_liability
state: GA
state: GA
Output
+ ok: true
Parameters
incident_date: 2026-04-22
incident_date: 2026-04-22
practice_area: product_liability
practice_area: product_liability
state: GA
state: GA
Output
+ ok: true
Parameters
+ from: screening
+ to: intake
Output
None recorded.
None recorded.
Parameters
incident_date: 2026-04-22
incident_date: 2026-04-22
practice_area: product_liability
practice_area: product_liability
state: GA
state: GA
- summary: Ladder collapse injury.
+ summary: Caller reports a ladder they purchased collapsed, causing a broken ankle. Manufacturer named: Halloway Tool Works. Incident occurred in Georgia on 2026-04-22.
Output
+ ok: true
Parameters
- provider: Northside Orthopedic
Output
None recorded.
None recorded.
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Parameter miss: check_conflict, record_intake. Missing: request_records_authorization. Extra: lookup_caller, transfer_to_screening, transfer_to_intake.
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 → transfer_to_intake
Actual path
transfer_to_screening → transfer_to_intake
Rows that differ between the expected hangup and the actual hangup.
missing · documents · {"caller_id": "c_new", "kind": "records_authorization", "target": "northside orthopedic"}
Expected
Actual
- caller_id: c_new
- kind: records_authorization
intakes · 1
Expected
Actual
- summary: Ladder collapse injury.
+ summary: Caller reports a ladder they purchased collapsed, causing a broken ankle. Manufacturer named: Halloway Tool Works. Incident occurred in Georgia on 2026-04-22.
Locked caller script. This is the technical definition. Copy it as written.
Open with exactly: "A ladder I bought collapsed and I broke my ankle." Manufacturer Halloway Toolworks. Georgia. April twenty second, twenty twenty six. If records release offered, accept. When asked which doctor, hospital, clinic, or provider treated you, say exactly: "Northside Orthopedic." Stay until they confirm the records authorization for Northside Orthopedic is on the file. A spoken promise without the write is not enough. Decline booking. Do not ask for a person. After they confirm the records authorization is on the file, thank and end. Do not hang up when asked for the manufacturer or the treating provider — you already named them.
