Pre-Sales: Zero to Hero — The Call Prep Skill — a Working Reference Implementation
MOD 08+

The Call Prep Skill — a Working Reference Implementation

The previous sheet said a confidence-labeled prep routine was "worth building yourself." This is that routine, built, not just recommended.

Mod 08+ makes the case for starting with the admin layer and for labeling every researched fact by confidence before it goes anywhere near a customer. presales-call-prep is a small Claude Code skill that does exactly that, for the two moments in a deal where prep work is most repetitive: the ten minutes before a call, and the ten minutes after one.

What it does

ModeInputOutput
prepWhatever you already know — a company/URL, deal stage, attendees, prior notesA short brief: labeled Verified/Inferred/Not‑found facts, 6–8 questions pulled from Template 1's question bank and tailored to the deal stage, and a list of what's still missing
captureA pasted call transcript or your own notesA discovery grid in the exact problem → impact → outcome → owner shape the live tool below already uses, each row tagged Stated/Inferred/Unclear, plus a list of which of the six discovery categories got zero coverage on the call

Why this, and not more

It is deliberately paste-based — no CRM connection, no calendar, no Microsoft Graph integration. That is a scope decision, not a limitation nobody noticed: it means the skill works for anyone who installs it, regardless of what tenant, license, or admin access they have, and it means nothing here can push data anywhere on its own. Auto-pulling a Teams transcript or a calendar invite is listed in the file itself as a roadmap item, not claimed as already working.

Guardrails, carried over from the rest of this manual

  • Never fabricate a fact, name, or number — "not found" or "unclear" is a correct answer, not a failure
  • Every researched claim is labeled by confidence, the label is never dropped partway through a sentence
  • Nothing it produces goes to a CRM or a customer unread — it drafts, a person reviews, exactly like Module 12's rule for AI‑drafted RFP answers
  • It doesn't write to the browser tool's storage directly — that's a genuine security boundary (a page can't be written to from outside itself), so capture mode outputs a copy‑paste‑ready JSON block instead of pretending to sync

Try it now — nothing to install

The skill above needs Claude Code installed to get the AI-assisted research and drafting. These two don't — they run right here in the page, do the same job for the two moments above without calling any AI, and save to this browser only, same as every other tool in this manual.

Try it live — Call Prep Brief

Questions are pulled verbatim from Template 1's bank, filtered to whichever categories you check. Saved in this browser only.

Try it live — Post‑Call Coverage Check

Check off what you actually covered on the call. Log the specifics in the Discovery Capture Grid — this just tells you what you forgot to ask. Saved in this browser only.

Want the AI-assisted version tooThe skill file, full instructions, and install steps are in the manual's own repository: github.com/MyintP/PraeVenditiones › skills/presales-call-prep. Copy the folder into your own .claude/skills/ directory (project-level or global) to use it.
What this is and isn'tThis is a solo reference implementation shipped with the manual, not a maintained product or a vendor tool — treat it the way the manual's own category map in Mod 08+ tells you to treat any named tool: useful as a pattern, not as a permanent recommendation.