AEmeddic-capture · intermediate
Auto-fill MEDDIC from Gong call transcripts
Last reviewed: 2026-05-23 · saves ~15 min/call/run
Our take
MEDDIC fields in Salesforce are notoriously rarely filled. AEs hate the admin. Result: deal-review meetings degenerate into anecdote. AI extracts MEDDIC from Gong transcripts automatically. Quality is ~85% — AE corrects the rest in 2 minutes.
Tool stack
Steps
- After every Gong call on a deal, fire a workflow.
- Workflow extracts: Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion.
- Each field is written back to the corresponding Salesforce custom field.
- AE gets a Slack notification: 'MEDDIC updated, please review.'
- Manager dashboards show MEDDIC completeness across portfolio.
Prompts
Extract MEDDIC from Gong call transcript · Claude Sonnet 4.6
You are a sales operations specialist. Given a Gong call transcript for opportunity {OPPORTUNITY_ID}, extract MEDDIC fields. For each:
M - Metrics: what business metric will success move? cite quote.
E - Economic Buyer: who has budget authority? cite quote or note "unclear".
D - Decision Criteria: what criteria will they use to choose? cite quote.
D - Decision Process: how + when will they decide? cite quote.
I - Identify Pain: what specific pain are they trying to solve? cite quote.
C - Champion: internal advocate, with title + behavior signal. cite quote.
For each field also output:
- Confidence: high / medium / low
- Evidence: direct quote from transcript with speaker name
Output JSON. If you can't fill a field, output {value: null, confidence: "low", evidence: "not discussed in this call"}.Pitfalls
- Don't fully autofill — AE review is the QA layer. Mark fields as 'AI-suggested' until AE confirms.
- Decision Process is the hardest field; AI often gets it wrong. Force a confidence score per field.
- If a Gong call has multiple deals discussed, AI can mix them up. Tag each call to a specific opp.