QA on every call, not 2%
Merdial scores 100% of calls against each client’s own rubric about four seconds after hang-up, coaches every agent live, and shows team leads the floor as it happens.
- 100%Calls scored, not sampled
- ~4sHang-up to a scored call
- ~300msObjection to answer
- Maya ChenOn a call06:41
“…so the audit flagged the log files, and that’s the part I’d fix first.”
Listen inWhisper - Jack HaleOn a call02:15
- Priya RamanCall endedScoring
- Leo MartinsAvailableReady
Manual QA in a call center means analysts listening to a small sample of calls, days after they happened. Automated call center QA scores every call as soon as it ends, so coaching, compliance checks and client reporting cover the whole floor instead of a sample.
- Opening and disclosure20/20
- Needs discovery24/30
- Accurate information25/25
- Conduct and tone11/15
- Clear next step10/10
Each client’s scorecard, on every call.
From a sample to 100%
- Weighted rubrics, QA bands, compliance and conduct flags, disputes and calibration
- Card numbers and identifiers are stripped before any generative model sees the transcript
- Scores arrive about four seconds after hang-up
Coach every seat at once
- Live coaching cards on every agent’s screen
- Live Calls: who is on a call, listen in, whisper
- A Coaching Queue and a Leaderboard for team leads
Where Merdial fits a contact center
- Outbound sales and appointment setting
- Floors that need a different standard per client
- Live coaching without a 50-agent minimum
- No SOC 2 report today
- No Linux desktop client — Windows 10+ and macOS 14.2+
Calls a floor handles hundreds of times a day
Editorial examples to show the idea. Your Merdial library is built from your own calls and approved by your own team.
- Objection
“Is this a sales call?”
They want honesty before they give time.
Reply shapeSay yes, say why you are calling in one sentence, and ask for thirty seconds.
- Objection
“Take me off your list.”
Not an objection — a request.
Reply shapeConfirm it and end the call. A rubric can check the agent did exactly that.
- Objection
“I already told the last person no.”
Fatigue from repeated calls.
Reply shapeApologise, acknowledge it, and do not pitch.
- Objection
“We’re under contract.”
A date, not a no.
Reply shapeAsk when it renews and whether you can follow up then.
Straight answers
How does Merdial compare to Balto, Cresta or Observe.AI?
Those tools also coach live, and at the top end they are serious competitors. They are also priced for roughly 50 to 100+ agents on annual contracts. Merdial’s advantage is the floor, not the ceiling: it starts at one seat, month to month.
Does Merdial have a SOC 2 report?
Not today, and the Security page says so. Tenants are isolated at the database level by row-level security, and card numbers and identifiers are redacted before any generative AI call.
How is pricing for 11 or more seats?
Custom plans are quoted to your volume: unlimited seats and accounts, talk time to your volume, a custom domain and branded login, SSO and a signed DPA.
What happens to card numbers said out loud on a call?
They are stripped, along with other personal identifiers, before the transcript reaches any generative model. That redaction is not optional and cannot be turned off to simplify a prompt.
Related
- CompareMerdial vs BaltoBalto starts at ~50 agents. Where it is the better buy, and where Merdial is.
- CompareMerdial vs Observe.AIObserve.AI starts at 100 seats. Where it is the better buy, and where Merdial is.
- FeaturesLive call monitoringWho is on a call right now, the live transcript, and listen-in and whisper controls for managers.
