Coach every cold call.
Sonar listens to every appointment-setter call your team makes, grades it across eight dimensions, and writes the next coaching session for you — with the transcript moment that earned each score.
“Just need 30 seconds — you get these calls all the time, so if it's not relevant tell me to do one.”setter
Not “how was the call.” Eight things you can teach.
Permission
score 3 · conf 83%Setter explicitly names a time constraint, states the topic, and offers a polite out. Textbook permission. The only gap is no active 'yes' from the prospect.
“Just need 30 seconds — if it's not relevant tell me to do one.”0:17
Press play. See what Sonar sees.
- 0:03setter
[brief audio artefact]
- 0:06setter
Hi there — quick one. Just saw you online, looks like you've built a strong local presence over the past few years.
- 0:12setter
You based in [city]? Yep, thought so.
- 0:17setter
Just need 30 seconds — you get these calls all the time, so if it's not relevant tell me to do one. Wanted to ask how you're getting new enquiries in at the moment.
- 0:28prospect
Honestly, not looking to change anything right now. Thanks though.
- 0:32setter
Totally fair. A lot of folks are in the same boat. Most of your work — coming in online, or more word of mouth?
- 0:39prospect
Word of mouth, mostly. I'm sorted. Thanks.
- Opening—
- Permission—
- Value clarity—
- Qualifying questions—
- Listening—
- Objection handling—
- Tone confidence—
- Next step—
Sample call · anonymised · real product output for every call your team makes
Real frameworks. Not just a model.
Most call-scoring tools are an LLM with a system prompt. Sonar grades against named, versioned, human-written rubrics — and lets the people on the calls argue back.
A versioned in-house rubric
v6.xDefines the eight scored dimensions and the language that separates a 2 from a 3. When the rubric updates, every historical call is re-scored against the new version automatically.
SPIN-aware qualifying analysis
Rackham · 1988Situation → Problem → Implication → Needs-payoff. Sonar checks whether the setter actually moved the prospect through the four question types — not just whether they asked questions.
Confidence on every score
always shippedEach score carries a model-confidence percentage. Below 70% the call surfaces in the human-review queue. The uncertainty is shown, not hidden.
Notes route to leadership
human-in-the-loopSetters disagree with scores. Coaches override them. The model learns from the conversation. The scorecard is a draft, not a verdict.
Coach the right rep, on the right thing. First.
Sonar ranks every rep on every dimension. The coaching queue tells you who to sit with this week, and which of the eight things to work on with each of them.
Most of the team clusters in the 0–2 band. The point of Sonar is to move the histogram right, one coachable habit at a time.
- Not interested36439%
- Other13615%
- Callback requested13014%
- Wrong number10011%
- Gatekeeper718%
- Voicemail536%
- Interested, no booking414%
- Meeting booked152%
One number. No tier games.
Unmetered calls. Unlimited reviews. Same number whether you have three reps or thirty.
Join the waitlist- Every call scored across 8 dimensions
- Per-dimension confidence percentages
- Strength + next-call focus per call
- Coaching queue + leaderboard + drift
- Versioned rubric · re-scored on update
- SPIN-aware qualifying analysis
- Notes route to leadership
- UK / EEA processing · DPA on request
Things we’re asked.
An in-house cold-call rubric refined over thousands of calls. Every dimension has a versioned rubric (currently v6.x) — when it updates, every historical call is re-scored automatically.
Hand your next coaching session to Sonar.
We’re onboarding teams in small batches. Tell us about your setup and we’ll be in touch when there’s a slot.