Sample report

What a result looks like

The agent said wording the policy prohibits. A potential violation always carries the quote it is based on, with its timestamp when the transcript has times; a reviewer decides what it means.

Result

Potential violation   QA score 70 / 100

Call acc-v1 · fictional voice agent · 1:15 · policy “Starter voice-agent policy”. Transcript provided with speaker labels.

Rules checked

Recording / consent disclosureconsent_disclosurehighNo issue

Recording/consent disclosure found in agent turn 1.

“This call is recorded for quality purposes.” at 0:00–0:04
Prohibited claimsprohibited_claimhighPotential violation

Agent speech contains prohibited wording: "we guarantee".

“We guarantee a full refund within two days, no matter what the contract says.” at 0:10–0:14
Sensitive data exposurepii_exposurecriticalNo issue

No sensitive data patterns detected in the transcript.

Script adherencescript_adherencemediumNo issue

All 2 script step(s) were found.

“Hello, thank you for calling Northwind Energy.” at 0:00–0:04
“Thank you for calling, have a great day.” at 0:30–0:34
Failed escalation / handofffailed_escalationhighNo issue

No request to speak to a human was detected.

Promised actionfailed_promised_actionmediumNo issue

No follow-up promises detected in agent speech.

Required closing wordingrequired_phraselowNo issue

Required wording found (1 of 4).

“Is there anything else I can help you with today?” at 0:20–0:24
Call completioncall_completionmediumNo issue

No failure signal in the available call data.

Transcript

  1. 0:00AgentHello, thank you for calling Northwind Energy. This call is recorded for quality purposes. How can I help you today?Shows: Recording / consent disclosure, Script adherence
  2. 0:05CustomerWill I get my money back for the cancelled installation?
  3. 0:10AgentYes. We guarantee a full refund within two days, no matter what the contract says.Evidence for: Prohibited claims
  4. 0:15CustomerGreat, thanks.
  5. 0:20AgentIs there anything else I can help you with today?Shows: Required closing wording
  6. 0:25CustomerNo, that is all. Thank you.
  7. 0:30AgentThank you for calling, have a great day.Shows: Script adherence

How to read this

  • This is post-call Voice QA: it evaluates calls that have already happened. It is not a simulation or pre-deployment test of an agent.
  • “Review recommended” is not a proven violation: it means a person should look, usually because something could not be confirmed.
  • Speaker labels matter: without them, checks that depend on who spoke can only be reviews, never passes or violations.
  • Quotations and timestamps are shown where the evidence supports them; some findings (for example a call that looks incomplete) have none.

Limitations

  • This is a QA signal for human review. It is not a compliance certificate, an audit opinion or legal advice, and it does not establish that the call was or was not lawful.
  • The result depends on the policy and wording rules listed here, on the transcript, and (for recordings) on automatic speech recognition, which can be wrong. Rules that depend on who spoke report “Review recommended” when speakers are unknown.
  • Evidence quotes are copied from the transcript. A rule that found something missing cites the stretch of the call that was examined, not the missing wording.
  • Disclosures made outside the call (an announcement before it, a website, a contract) are not visible to this check.

Results are QA signals for human review. Every quoted line above is real transcript text; the policy that produced them is editable per organization. Read how to read a result, or start free and check your own calls.

Sample report · GoldNovaAI