Your booking agent has to get the address right.
Dispatch calls turn on the exact street, the exact window, and a quote the company can actually honor. One misheard address and the truck rolls to the wrong house. Roark scores every call — on the audio.
Backed by YCCaller: It's 80 Maple, not 18.
Agent: Got it — sending the tech to 18 Maple.
Misheard address — Street number flattened by the ASR — heard by the audio model, invisible to the transcript.Address capture
Scoring production voice AI for teams at
§01 · When the call goes wrong
Here's how a booking call goes wrong.
Each one is a wasted truck roll, a no-show window, or a quote you have to eat at the door — and most are inaudible to a tool that only reads the transcript.
The 'Eighty' vs 'Eighteen' address
Caller: It's 80 Maple, not 18.
Agent: Got it — sending the tech to 18 Maple.
The transcript reads clean; the audio is wrong. Only an audio model hears the digit the ASR flattened — and that truck rolls to the wrong house.
Address captureThe wrong arrival window
Caller: Morning only — I work after noon.
Agent: Perfect, you’re booked 2–4pm.
The constraint was stated and ignored. Roark fails any booking where the confirmed window contradicts what the caller asked for.
Window accuracyThe price you can’t honor
Caller: So how much for the drain clear?
Agent: That’ll be a flat $89, all in.
A number invented on the call that no price book backs. The customer holds you to it at the door — Roark flags any quote the agent can’t source.
Quote integrityDead air on the emergency call
Caller: My basement is flooding right now.
Agent: …
A four-second silence on a 2am burst pipe and the caller hangs up and dials a competitor. Roark scores response time and dead air the transcript never shows.
Response timeThe pushy add-on tone
Caller: I just need the one repair, thanks.
Agent: You really should add the $400 membership today.
The words might pass; the delivery is aggressive. The audio model scores tone and pressure the transcript reads as a polite suggestion.
Tone§02 · From caught to fixed
Roark catches every one of these — and proves the fix.
Each failure above is filed with its evidence, becomes a repeatable simulation until a candidate passes, and is verified on your next thousand live calls.
Your fix, replayed against the exact failures above.
Every change explicit and diffed — you apply it.
You ship — Roark confirms the metric moved on live calls.
you ship it — Roark verifies every call, same booking flow, every channel
…and the loop runs again on the next call.
§03 · Simulate before launch
Break it in staging,
not in production.
Run your agent against hundreds of simulated callers — realistic personas, accents, background noise and edge cases — and get every conversation scored before a customer ever dials in.
Scenarios & personas
Hundreds of simulated callers — the angry one, the rambler, the interrupter — built from your real call types.
45 languages & accents
Native accents, code-switching and background noise — in every market your agent answers.
Load & health tests
Peak-volume concurrency and always-on health checks, so the agent that passed in staging survives launch day.
Run it in CI
Every prompt or model change runs the suite before it merges — quality gates for conversations, not just code.
1 failure filed as an issue — fix it before launch, not after
§04 · Evals & observability
64+ metrics. Your models,
not just an LLM.
Every production call scored as it lands — issues filed, alerts fired, dashboards and OTEL traces on tap, for voice calls and chat threads alike. And where most tools grade a transcript with an LLM, Roark runs purpose-built audio models on the call itself, measuring what your customer actually heard.
Everyone else
LLM reads the transcript
“The agent said the right words.” Misses how it sounded — the mispronounced drug name, the flat apology, the rushed close.
Audio models hear the call
Pronunciation, accent, emotion and vocal stress measured from the waveform — the signal an LLM grading text can never see.
Audio-native
custom models
- Address capture
- Accent clarity
- Pronunciation
- Dead air
- Vocal stress
- Interruptions
Booking accuracy
task
- Window accuracy
- Quote integrity
- Service-type match
- Slot confirmation
Conversational
LLM + rules
- Tone
- Task success
- Hallucination
- Repetition
- Upsell pressure
- Empathy
Performance
latency
- Response time
- Time-to-first-word
- ASR WER
- Barge-in handling
§05 · Get started
First call scored in under a minute.
One click on any platform below and production calls stream in on their own — or send any recording with three lines of code.
Read the quickstartimport Roark from '@roarkhq/sdk'const roark = new Roark({ apiKey })await roark.calls.evaluate({recordingUrl, agent: 'support_v2',}) // scored in seconds
Works with
Also built for
Customer Support
Resolve it for real, escalate cleanly, and hear the frustration the transcript hides.
ExploreRetail
Quote the policy as written, never invent stock, and hold your brand voice when a returns caller is furious.
ExploreInsurance
Quote the coverage right, read every required disclosure, and never sound cold to a claimant in crisis.
Explore
HIPAABAA availablePayment & call-recording controls
Roark checks card and PCI-sensitive turns are handled to script and flags recording-disclosure gaps — with configurable retention on every call.
Bring a recording.
We’ll score it live.
See your own agent measured on the audio it actually produced — in the demo, in real time. Stop guessing whether your voice AI works.
founders@roark.ai · we reply fast