meeting-prep-brief

# Meeting Prep Brief

Generate a concise pre-call intelligence brief in 2-3 minutes. Everything you need before walking into a meeting: company context, stakeholder profile, competitive landmines, talking points, and a proposed agenda. Uses WebSearch for real-time intelligence and builds on existing account research when available.

## Steps

### 1. Collect Input

Ask the user for the following:

- **Company name** (required)
- **Contact name and title** (required) — who you are meeting with
- **Meeting type** (required) — one of: discovery call, demo, follow-up, QBR, negotiation
- **Your product/service** (optional) — if provided, talking points and agenda will be tailored to the user's offering
- **Deal context** (optional) — past conversations, where they are in the funnel, previous meetings, known objections
- **Specific questions** (optional) — anything the user wants answered or addressed during the meeting

Do NOT proceed until you have the company name, contact name/title, and meeting type.

### 2. Check Existing Intelligence

Use Glob to check for existing files that can accelerate the brief:

- `docs/accounts/{company}.md` — account research brief (from the Account Research Brief skill)
- `docs/signals/{company}.md` — signal data (from the Signal Scanner skill)
- `docs/pipeline/scored-leads.md` — qualification score
- `docs/icp.md` — ICP match assessment

**If any of these exist:** Read them. Summarize what you loaded. Skip redundant research in subsequent steps and build on the existing intelligence instead.

**If none exist:** Proceed with full research from Step 3 onward.

### 3. Company Context

Skip this step if a recent account brief already exists at `docs/accounts/{company}.md`.

Run 2-3 WebSearch queries (e.g., `"{company} what does it do"`, `"{company} news 2026"`, `"{company} funding"` ).

Gather:

- What they do (one paragraph, plain language)
- Recent news from the last 30 days (1-3 headlines with source URLs)
- Key metrics: funding stage, total raised, headcount, revenue range (if public)
- Current challenges or priorities — inferred from press releases, job postings, executive interviews, or earnings calls

If any data point is not findable, write "Not found." NEVER fabricate company data.

### 4. Stakeholder Profile

Run 2-3 WebSearch queries focused on the specific contact (e.g., `"{contact name}" "{company}"`, `"{contact name}" LinkedIn`, `"{contact name}" podcast OR conference OR interview`).

Gather:

- **Current role and tenure** — title, how long they have been in this position
- **Career history** — where they worked before, what they did (1-2 sentences)
- **Public content** — LinkedIn articles, podcast appearances, conference talks, blog posts, tweets. Include URLs where found.
- **Communication style clues** — are they technical or business-oriented? Data-driven or narrative-driven? Formal or casual? Infer from their content and background.
- **What they likely care about** — based on their role and company context, what metrics, outcomes, and problems occupy their day
- **Potential common ground** — shared background, interests, mutual connections, or overlapping communities

**Rules for this step:**
- If WebSearch returns nothing about the contact, say "No public profile found for {contact name}. Recommend checking LinkedIn directly before the call." Do NOT fabricate biographical details.
- Do not guess at personality traits without evidence. Only note communication style clues when backed by observed content.

### 5. Competitive Landmines

Run 2-3 WebSearch queries (e.g., `"{company}" uses "{competitor tool}"`, `"{company}" tech stack job postings`, `"{company}" case study "{competitor}"`).

Gather:

- **Current tools/solutions** in the user's product category — detected from job postings, integrations pages, case studies, G2 reviews, or tech stack analysis tools
- **Competitor relationships** — are they an existing customer of a competitor? Have they published a case study or testimonial with them? Are they listed on a competitor's customer page?
- **Potential objections** based on the competitive situation (e.g., switching costs, existing contract, "good enough" inertia)
- **Positioning recommendations** — 1-2 sentences on how to position against the incumbent (if one is identified) or how to frame the conversation if no competitor is in place

If the user did not provide their product/service, keep this section general: focus on what solutions the company uses in their operations, and flag any vendor relationships that could be relevant.

### 6. Meeting-Type Specific Prep

Generate questions and preparation items based on the meeting type the user specified.

#### For Discovery Calls

- **5 open-ended questions** to understand their current state (process, tools, team structure, priorities)
- **3 pain-probing questions** based on likely challenges inferred from the company context and stakeholder profile
- **2 vision questions** ("Where do you want to be in 12 months?" style)
- **Red flags to watch for** — 3-4 signals during the call that indicate this prospect is not a fit (e.g., no budget allocated, evaluating in 12+ months, no pain acknowledged)

#### For Demos

- **3 most relevant features/capabilities to show** — based on the contact's likely pain points and company context
- **Potential "wow moment"** — one thing to lead with that will immediately resonate
- **3 objections to prepare for** — based on competitive landscape and company situation
- **1-2 success stories** — describe the type of similar company/use case to reference (e.g., "a Series B fintech that cut onboarding time by 40%"). Do not fabricate specific customer names.

#### For Follow-ups

- **Previous context summary** — restate what the user provided about past conversations
- **3 questions that should have been asked** — gaps to fill based on what a thorough discovery would cover
- **New information since last meeting** — any recent signals (funding, hires, news) discovered