Score every sales call and send each rep daily coaching
Every weekday evening, each recorded call gets scored against your QA rubric, logged for managers, and turned into a private coaching note for the rep.
Cartesia
Aircall
Google Sheets
SlackSalesOperationsAI ReportsNotifications & AlertsEvery weekday at 6pm, run a call quality review on my team's Aircall conversations and give each rep a scored, coachable write-up. Use a cron trigger so the run covers a single closed day and never scores the same call twice.
Start with Aircall List Calls to pull today's calls, using the from and to filters (they take UNIX timestamps in seconds) to bound the window to today only, and page through results until the day is complete. Keep only calls that were actually answered, that have a recording URL on the call record, and that ran longer than two minutes. Aircall returns a recording URL only for calls that were genuinely recorded, so if a call has no recording, skip it quietly and move on rather than failing the whole run. Missed calls, voicemails, and very short calls are not worth scoring. Aircall allows 120 requests per minute per company, so pace the run and avoid hammering the API when the day has a lot of calls.
For each remaining call, download the recording and send the audio to Cartesia Transcribe Audio (STT), which is the batch Ink Whisper speech to text endpoint. The audio goes as a multipart file part, and accepted formats include mp3, wav, flac, ogg, and webm. Every Cartesia request must include the Cartesia-Version header pinning the API date, or it will be rejected. If a transcript comes back empty or the transcription fails for one call, log that and continue with the other calls.
Score every transcript against a fixed rubric. Let me paste my own criteria if I have a scorecard my team already uses, and otherwise default to these five: opening and agenda, discovery questions asked, objection handling, clear next step booked, and tone and compliance. Give each criterion a score from 1 to 5, and for each one include a verbatim quote pulled straight from the call transcript as evidence for the score, plus a coaching note of about two sentences saying what the rep should do differently next time. Add up the five criteria into a total out of 25. Be specific and fair: quote what the rep actually said rather than paraphrasing, and if a criterion genuinely did not apply on that call, say so instead of inventing a score.
Append one row per scored call to my Google Sheets QA log using Append Values, so managers can trend performance over time. Ask me for the spreadsheet and tab to use. Each row should carry the call date, the rep's name, the customer or phone number, the call duration, each individual criterion score, the total out of 25, and the overall coaching note. Keep the column order identical on every run so the sheet stays chartable.
Attach the result back to the call record in Aircall. Use Comment a Call to post the scorecard on the call itself, showing each criterion score, the total out of 25, and the coaching note, so the score lives next to the conversation it describes. Then use Tag a Call to tag any call scoring under 15 out of 25 as coaching-needed, so weak calls can be filtered later.
Finally, message each rep privately in Slack. Resolve the rep to their Slack user (List Users can match on email, and Aircall exposes the agent's email on the call record), then use Open a Conversation to get the direct message channel ID and Send a Message to deliver the note. Keep the DM short and constructive: the call it refers to, their single weakest criterion, the quote that earned that score, and the coaching note. If a rep had several calls scored that day, send them one combined message rather than a stream of separate DMs.
Escalate sparingly. Only notify the sales manager when a call scores under 10 out of 25 or when the transcript contains a compliance miss, such as a missing recording disclosure or a promise the rep is not allowed to make. Ask me who the sales manager is and DM them the rep name, the call, the total score, and the specific reason it was escalated. Everything else stays between the workflow and the rep.
Keep the scope on scoring and coaching. Do not write call summaries into a CRM, and do not use this to build an account-level recap. If there are no qualifying calls on a given day, do nothing and stay quiet.
What does this prompt do?
- Reviews every substantial call your team took that day, not just the handful a manager had time to listen to, and scores each one against a five part quality rubric covering the opening, discovery, objection handling, next steps, and tone.
- Backs every score with a real quote from the conversation, so a rep can see exactly which moment earned the rating instead of arguing with a number.
- Keeps a running quality log in a spreadsheet with one row per call, so managers can trend performance by rep and by criterion over weeks rather than guessing.
- Writes the score onto the call record itself and flags weak calls for coaching, then sends each rep a private message about their single weakest area with a short, specific note on what to do differently.
What do I need to use this?
- An Aircall account with call recording already switched on, since only recorded calls can be reviewed
- A Cartesia account, used to turn each recording into a written transcript
- A Google account with a spreadsheet where the quality log should be kept
- A Slack workspace so reps and the sales manager can receive their notes privately
- Your team's quality rubric if you already use one, plus the name of the manager who should receive escalations
How can I customize it?
- Swap in your own rubric criteria. The five defaults are a starting point, and most teams paste in the scorecard they already use for call reviews.
- Change the timing and the call filter, for example reviewing only calls longer than five minutes, only outbound calls, or running at lunchtime instead of end of day.
- Move the thresholds that decide when a call gets flagged for coaching and when it gets escalated to the sales manager, so the volume of alerts matches how closely you want to supervise.
FAQs
Do reps see their own scores?
Does this replace my sales manager listening to calls?
What happens to calls that were not recorded?
Will the scores be consistent from day to day?
Do I need to tell my team their calls are being reviewed?
Related templates
Stop buying stale lists. Reps run a saved search, work the results like an inbox, and only the accounts they approve ever reach your CRM.
Drag creators through Sourced to Wrapped on a board grouped by campaign, with audience stats on every card and a one-click brief for each creator.
Search creators for free, shortlist the best with your team, and spend an audit credit only when you decide someone is worth a closer look.
Pick a repricing rule, send an assistant out to check competitor pages, then approve the new prices that clear your margin floor.
See the projected bounce rate for any outbound send before it goes out, and keep the launch button locked until the list is clean enough to be safe.
One screen showing invited, completed, and shortlisted counts for every open role, plus the stalled invites that have quietly gone nowhere.
Every call reviewed, not just the ones a manager had time for.
Give every rep specific, quoted feedback on their calls each day, and give managers a quality trend they can actually act on.