Call analysis and review with AI
AI Sales 11 min
0/19
Business problem
The call ends and... chaos begins: notes in 3 places, a follow-up written from memory 2 days later, no analysis of 'what could I have done better'. Without analysis every call is a one-off event — you repeat the same mistakes and don't extract the signals that were in the call.
Proven method
After the call (transcription — with client consent or from notes):
- AI GENERATES: call summary, customer pain (what specifically hurts), needs (what they need), objections (which came up), buying signals (what indicates interest), competitors (what they mentioned), next steps (agreed and missing), deal risk (1-10 assessment + why),
- AI CALL REVIEW: what I did well (examples from the transcription), what badly, what I didn't ask (discovery gaps), where I lost control of the call (moments when I was being led instead of leading),
- NEXT STEPS: AI prepares a follow-up with specifics from the call. Rules: transcription only with consent (say 'I'm recording so I can prepare a better follow-up'), deal risk isn't a verdict — it's a risk list to address, call review is for yourself (not for your boss) — it has to be brutally honest.
Market example
- BAD
- Improvised general pitch lacking buyer context or verified data.
- GOOD
- After a call with an e-commerce COO AI returns: SUMMARY: 'the company is growing, inquiry handling 24h+, a 2-person team without tools'. PAIN: 'response time, no escalation in the season'. NEEDS: 'a queue system, SLA, email integration'. OBJECTIONS: 'a concern about cost, a question about implementation time'. BUYING SIGNALS: 'asked about prices and deadlines, mentioned the season'. DEAL RISK: 7/10 — 'budget unconfirmed, no authority (the COO must agree with the CEO)'. CALL REVIEW: 'good — I asked about the process; bad — I didn't clarify the budget; I didn't ask about the CEO's decision; I lost control at the price question — I went into negotiations too early'. FOLLOW-UP: 'thanks for the call, summary: problem X, I propose a demo on Wednesday'.
Ready-to-use prompt
You are a sales call analyst. Context: I have a transcription/notes from a call: [PASTE THE CALL OR NOTES]. I sell [PRODUCT]. Task: Do the analysis: 1. Call summary (5 sentences) 2. Customer pain (specifically, with quotes) 3. Needs (what they need) 4. Objections (which and how I addressed them) 5. Buying signals (what indicated interest) 6. Competitors (what they mentioned) 7. Next steps (agreed vs missing) 8. Deal risk (1-10 + a risk list) 9. Call review: what I did well / what badly / what I didn't ask / where I lost control (with examples) 10. A ready follow-up (max 100 words) Format: 10 sections.
Model rekomendowany: Claude 3.5 Sonnet / GPT-4o Copy and paste into ChatGPT, Claude or Perplexity
ALTERNATYWNE SCENARIUSZE
Variations & alternative scenarios
No transcription version: analysis from notes (AI proposes what to add after every call).
Team version: a standard analysis template + sharing conclusions.
Practical exercise
Take your last call (or record a test one with AI). Generate the analysis in 10 sections. Write 2 things you'll change in the next call based on the call review.
Zapisano w przeglądarce