call-prep

name	call-prep
description	Prepare a pre-meeting brief for a sales call — account snapshot, attendee persona mapping, matched case study, likely objections with rebuttals, sharp discovery questions, and a clear ask. This skill should be used when the user says "prep me for my call with [company]", "I'm meeting with [company] prep me", "call prep for [person]", "what should I know before talking to [company]", "prepare for my meeting with [X]", or "brief me on [company] before my call". Reads PROFILE for markets and ICP weighting. Produces a one-page brief saved to the account folder (`content/<active>/accounts/<account-slug>/`).
metadata	{"version":"0.3.0","phase":"3","capability_tier":"core"}

GTM — Call Prep

Build a crisp pre-meeting brief the colleague can read in five minutes. Map the account to the ICP, map each attendee to a persona, select the closest proof story, arm the colleague with objection responses and discovery questions, and close with one clear ask. Read-only research only; never send anything.

Resolve the active profile (the agent provides it; everything loads from profiles/<active>/, never plugin/).

Load context (in this order)

Knowledge resolution (product-aware). Wherever this skill loads a per-product knowledge file — icp-personas.md or market-scan-config.md — resolve its path with python -m gtm_core.resolve_knowledge <file> --profile <active> [--product <slug>] and read whatever path it prints, instead of opening knowledge/<file> directly. The helper returns the product-level file (products/<slug>/<file>) when present and falls back to the profile-level knowledge/<file> otherwise. Pass --product when the run is bound to one product (the lead default_product from PROFILE.md, or a product the operator named); omit it for profile-wide work — a profile that keeps one shared knowledge pack always falls back to the profile level, so nothing changes for it.

1. PROFILE — profiles/<active>/PROFILE.md. Pull: name, brand_name, target_markets, icp_weighting, language. Respect language throughout.
2. profiles/<active>/knowledge/icp-personas.md — qualification criteria, scoring rubric, persona cards (Pain · Claim · Gain), buying committee roles, and "why now" triggers.
3. profiles/<active>/knowledge/case-studies.md — four shapes + case-study selection map. Pick the closest case by shape first, then industry.
4. profiles/<active>/knowledge/product.md — the product suite (resolve real product names from here; declared in PROFILE.md → products[]), solution themes, features, discovery questions.
5. profiles/<active>/knowledge/company.md — company narrative, team, investors, competitive positioning.
6. Prior prospect file (if it exists in the account folder content/<active>/accounts/<account-slug>/) — a prospects-*-[company].md or outreach-*-[company].md from a prior run. Pull the score, why-now signal, and persona details already captured. Skip the web research if the file is fresh (≤30 days).

Gather inputs

Ask the colleague:

• Company name (required).
• Attendees — names and titles if known; segment them into Enterprise / Startup.
• Meeting type — discovery, demo, follow-up, pilot review, QBR, partner call. Default: discovery.
• Stage — first touch, mid-cycle, late-stage. Default: first touch.
• Any context they want to use — recent news, a specific pain they mentioned, a prior email thread.

Keep the ask brief — one short message, not a form. If they've already given context in their request, use it and don't ask again.

Research the account (skip if prior file is fresh)

Run a focused web sweep. Do NOT use metered tools (Firecrawl, Vibe Prospecting) for call prep — free paths only (web search + browser).

Gather:

1. Company signals — recent job posts mentioning AI agents or LLM/agent frameworks (LangGraph, CrewAI, AutoGen, MCP, A2A), press releases, engineering blog posts, LinkedIn activity.
2. Agent maturity signals — are they building, piloting, or running agents in production? Evidence from open sources.
3. Regulatory exposure — HQ jurisdiction + industry → infer relevant regulators from profiles/<active>/knowledge/icp-personas.md (MAS, HKMA, SEC/FINRA, HIPAA/HSA, EU AI Act, PDPA, etc.).
4. Attendee LinkedIn — title, tenure, background, recent posts. Map to the six personas in profiles/<active>/knowledge/icp-personas.md. Infer primary pain from their role.
5. Recent news — funding rounds (past 18 months), product launches, security incidents, compliance announcements.

Stop after the top signals are clear. This is a 5-minute sweep, not a full prospect run.

Build the brief

Structure the output as follows. Keep each section tight.

1. Account snapshot (4–6 lines)

Company · segment (Enterprise / Startup) · industry · HQ + markets where they operate · employee count + recent funding · ICP score (apply the rubric from profiles/<active>/knowledge/icp-personas.md, show the number and the top 3 factors driving it).

2. Meeting context

Type · stage · attendees with role and persona mapping.

3. Attendee persona cards (one per attendee)

For each: Title → Persona → primary pain (from persona card) → what they care about proving in this meeting → how to open with them.

If multiple attendees, identify the economic buyer vs. champion vs. influencer.

4. The "why now" signal

The one real, specific, dated signal that makes this meeting timely. Use the prior prospect file if available; otherwise use the best signal from the web sweep. If no strong signal was found, say so — do not invent one.

5. Matched proof story

Apply the case-study selection map: shape first, then industry. Name the case study, one-sentence what they did, one-sentence outcome metric. Note the shape (internal multi-agent workforce / cross-org B2B exchange / machine-as-consumer commerce / multi-party verifiable workflow) and why it maps.

If the prospect has a regulatory angle, lead with the compliance-as-wedge framing from profiles/<active>/knowledge/case-studies.md.

6. Three likely objections + rebuttals

Objection	Rebuttal
(most likely for this persona)	(draw from profiles/<active>/knowledge/product.md competitive differentiation + profiles/<active>/knowledge/case-studies.md proof)
…	…
…	…

Pick objections relevant to the attendees' personas. Common ones: "We're already on [hyperscaler IAM]" / "We don't have agents in production yet" / "We can build this ourselves" / "We're too early-stage for this kind of overhead".

7. Discovery questions (5–7)

Pull from the discovery questions in profiles/<active>/knowledge/product.md; adapt to the meeting type and stage. For first-touch discovery, prioritise urgency-creating questions. For later-stage, prioritise technical/commercial fit.

8. The ask

One specific, low-commitment action to close this meeting with. Match to stage:

• Discovery → schedule a scoped technical deep-dive or PoC scoping call.
• Demo/follow-up → agree on a PoC success criterion or intro to the technical evaluator.
• Pilot review → next milestone agreement or commercial proposal.
• Partner/BD call → a named next step with a named person and date.

Never end a meeting without one.

Output

Save the brief as call-prep-[company]-[YYYY-MM-DD].md in the account folder content/<active>/accounts/<account-slug>/ (see CLAUDE.md "Per-account outputs").

Tell the colleague: "Your brief for [company] is saved as call-prep-[company]-[date].md. Here's the quick version:" — then paste the brief inline so they don't have to open the file first.

Guardrails

• Never use metered tools (Firecrawl, Vibe Prospecting) for call prep — free paths only.
• Never invent a "why now" signal. If the sweep finds nothing strong, say so and suggest the colleague bring their own signal into the meeting opener.
• Never fabricate attendee detail. If title is unknown, map by company stage and infer the most likely person (CEO for Startup first touch; Head of AI Platform for Enterprise first touch).
• Match proof stories by shape, not by vertical keyword alone.
• Keep the brief to one readable page — under 600 words of prose, tables OK.