Identify Skill Gaps From a Rep's Call Recordings

You are a sales manager doing a fair, data-grounded diagnosis of a rep's performance at a digital marketing agency. You diagnose skill gaps from call evidence — you do not coach in this output.
Agency: [AGENCY_NAME] — [SERVICES] | Rep: [REP_NAME] | Tenure: [TENURE_MONTHS] months | Call type: [CALL_TYPE]
Transcripts (or summaries) of [N_CALLS] calls:
[CALL_RECORDINGS]
Agency's expected discovery framework (if any): [DISCOVERY_FRAMEWORK]
Identify the top 3 recurring selling-skill gaps that show up across [REP_NAME]'s calls, ranked by frequency and revenue impact. For each, quote 1-2 specific lines from the transcripts as evidence.

- Evidence > opinion. Every gap must be supported by a direct quote with the call number.
- Distinguish skill gaps (how they sell) from product-knowledge gaps (what they don't know about [SERVICES]).
- Do not invent quotes. If a pattern only appears in one call, say so and lower confidence.
- Skip personality observations ("sounds nervous"). Focus on observable selling behaviors: discovery depth, objection handling, pricing posture, framing of value, next-step control.
- No coaching recommendations — that's the next stage.

Output:
1. Skill-gap summary table (Gap | Calls where it appeared | Confidence)
2. For each of the top 3 gaps: name → 1-2 quoted evidence lines → revenue impact hypothesis
3. Strengths observed (1-2 things the rep does well, for balance)
4. What you'd need more of to confirm

• [AGENCY_NAME] — Your agency
• [SERVICES] — What you sell
• [REP_NAME] — Rep being reviewed
• [TENURE_MONTHS] — Months in role
• [CALL_TYPE] — Discovery, proposal review, closing call, etc.
• [N_CALLS] — Number of calls being analyzed (3-5 ideal)
• [CALL_RECORDINGS] — Paste Gong/Fathom/Fireflies transcripts or summaries, labeled Call 1, Call 2, etc.
• [DISCOVERY_FRAMEWORK] — Your house framework: MEDDIC, SPIN, custom — or 'none'

Agency: Northwind — SEO + paid social | Rep: Marco | Tenure: 9 months | Call type: Discovery | N_CALLS: 3
Framework: MEDDIC.
Call 1 excerpt: Prospect: 'What does this cost?' Marco: 'Most clients are around $4-6k a month.' (no qualifier)
Call 2: Marco does not ask about decision process; books proposal anyway.
Call 3: Prospect mentions 'we tried an agency before and got burned.' Marco: 'Yeah, that happens — let me show you our process.'

Skill-gap summary:
| Gap | Calls | Confidence |
| Premature pricing without qualification | 1, 3 | High |
| Skips decision-process discovery | 1, 2 | High |
| Doesn't probe past-agency pain | 3 | Medium (single call) |

1. Premature pricing (High). Call 1: "Most clients are around $4-6k a month" given before scope or budget surfaced. Likely anchors low; predicts margin pressure at proposal.
2. Skips decision process (High). Calls 1 & 2: no MEDDIC "D" question. Predicts late-stage stalls when an unseen approver appears.
3. Surface-level on past-agency pain (Medium). Call 3 pivot to "our process" cuts off a high-signal disqualifier or differentiator moment.

Strengths: Builds rapport quickly; transitions to agenda cleanly in all 3.

Need more: 2-3 additional discoveries, ideally with deals that DID convert, to test whether the pattern persists in wins.

• Feed 3-5 calls minimum. One call = a moment; 5 = a pattern.
• Mix won and lost calls so Claude can compare behaviors.
• Paste the agency's house framework — without it Claude will default to generic SPIN/MEDDIC and miss your context.