Meeting Prep Agent

Meeting Prep Agent

Never walk into a meeting cold again. Give the Meeting Prep Agent a list of attendees and it researches each person across LinkedIn, Google News, and their company website, then uses AI to synthesize everything into a 2-minute briefing. The output is an original document built by orchestrating multiple Apify Actors and an LLM.

Powered by Apify Actors orchestration and OpenRouter AI (DeepSeek V4 Flash).

Why use Meeting Prep Agent?

• Sales and BD - Prepare for prospect calls without manual research
• Investors and consultants - Brief yourself before portfolio or client meetings
• Recruiters - Understand a candidate's background and recent activity before a screen
• Anyone with external meetings - Turn calendar invites into actionable context

The AI layer is load-bearing. It doesn't just collect data, it synthesises across sources into a structured briefing with talking points, open questions, and company context you would not get from any single source.

How it works

1. Enter the attendee name and company. LinkedIn URL and role are optional.
2. The Agent runs 3 research tasks in parallel:
   • LinkedIn Profile - Searches for the profile if no URL given, then extracts role, background, and career history
   • Google News - Finds recent mentions, interviews, and coverage
   • Company Website - Captures product positioning and description
3. AI synthesises everything into a structured briefing
4. Output is saved as JSON to the dataset and as Markdown to the key-value store

Input

Fields

Field	Type	Required	Description
attendeeName	string	yes	Full name of the person you're meeting with
company	string	yes	Company or organization the attendee works for
role	string	no	Job title or role (improves research accuracy)
linkedinUrl	string	no	LinkedIn profile URL (skips the Google search step)
meetingContext	string	no	What the meeting is about (improves relevance)
maxActorCostUSD	number	no	Cost cap for sub-Actor calls (0 = no limit)

If no LinkedIn URL is given, the Agent searches Google for the profile automatically.

Output

Dataset (JSON)

{
    "meetingSummary": "2-3 sentence overview of who you're meeting and why it matters",
    "attendees": [
        {
            "name": "Jana Dvorak",
            "role": "VP of Product",
            "company": "Productboard",
            "background": "3-4 sentences on career, expertise, relevant context",
            "recentActivity": ["notable recent posts, talks, interviews"],
            "likelyPriorities": ["inferred from role and company context"],
            "icebreakerAngle": "one genuine, non-cringey conversation starter"
        }
    ],
    "companySnapshot": {
        "name": "Productboard",
        "oneLiner": "Product management platform",
        "stage": "Series B",
        "recentNews": ["announced new AI features"],
        "inferredPriorities": ["what they are probably focused on right now"],
        "watchOuts": ["anything sensitive or worth being aware of"]
    },
    "suggestedTalkingPoints": ["string"],
    "openQuestions": ["things worth asking in the meeting"],
    "sources": ["https://linkedin.com/in/jana-dvorak", "https://..."]
}

Key-value store (Markdown)

A formatted Markdown version (briefing.md) is also saved to the key-value store with section headers, bold labels, and bullet lists. Readable directly in the Apify Console or downloadable.

Pricing / Cost estimation

Pay-per-run. Each run incurs costs from:

• Sub-Actor calls (LinkedIn scraper, Google search, website crawler) - typically $0.01-0.05 each
• OpenRouter AI inference (DeepSeek V4 Flash) - approximately $0.01-0.03 per briefing

Total cost per run: roughly $0.05-0.20 for a single attendee at quick depth. Use maxActorCostUSD to cap unexpected spend in case a sub-Actor runs longer than expected.

Tips

• Provide linkedinUrl when you have it to skip the Google search step
• Fill in meetingContext for significantly better briefing relevance
• Provide OPENROUTER_API_KEY via apify secrets add openrouterApiKey <key> or set it in the Apify Console env vars

Local development

# Edit the sample input with your OpenRouter key
# storage/key_value_stores/default/INPUT.json

# Run locally (storage/ is local-only)
apify run

Deployment

$apify push

FAQ, disclaimers, and support

• What if a sub-Actor fails? The Agent degrades gracefully. If LinkedIn is blocked, the briefing notes the gap and continues with Google News and website data.
• Is LinkedIn data always accurate? The Agent scrapes whatever is publicly visible. Results depend on the target profile's privacy settings and the scraper's capabilities.
• Does this comply with platform terms? You are responsible for ensuring your use of third-party scrapers and LLM APIs complies with their respective terms of service.
• Need help or want a custom version? Open an issue on the Issues tab or contact us directly.

Resources

• Apify SDK for JavaScript
• Apify Platform
• OpenRouter API
• LinkedIn Profile Scraper
• Google Search Scraper
• Website Content Crawler