Person Researcher

Researches a specific person's public activity and presence for hyperpersonalized outreach. Finds their LinkedIn posts, conference talks, articles, Twitter activity, career trajectory, and interests. The output feeds into message-generator Tier 3 personalization.

Prerequisites

• WebSearch tool available
• Person's name (required) and at least one of: company, title, LinkedIn URL, email domain
• Optional: agency.config.json for service context (to identify relevant conversation topics)

Phase 0: Intake

1. Read agency.config.json from the project root (if available).
2. Extract services[].name and icp.segments[].titles to know which topics are relevant to the agency's pitch.
3. Accept parameters:
   • name -- (required) full name of the person
   • company -- (recommended) their current company
   • title -- (optional) their job title
   • linkedin_url -- (optional) direct LinkedIn profile URL
   • email -- (optional) for additional search context
   • depth -- quick (3-5 searches) | standard (8-12 searches) | deep (15+ searches). Default: standard

Phase 1: LinkedIn Activity

WebSearch queries:

• "{{name}}" site:linkedin.com/in -- find their profile
• "{{name}}" "{{company}}" site:linkedin.com/posts -- find their posts
• "{{name}}" "{{company}}" site:linkedin.com/pulse -- find their articles

Extract:

• Profile URL: Their LinkedIn profile link
• Current title and company: Verify against provided data
• Recent posts (last 3-6 months):
  • Topic / subject of each post
  • Date (approximate)
  • Key points or opinions expressed
  • Engagement level (if visible in snippets)
• Articles published: LinkedIn Pulse articles or newsletter posts
• Activity themes: What topics do they post about most? (e.g., ecommerce trends, marketing, leadership, hiring)

Phase 2: Conference Talks and Podcast Appearances

WebSearch queries:

• "{{name}}" "{{company}}" "speaker" OR "keynote" OR "panelist" OR "conference"
• "{{name}}" "{{company}}" "podcast" OR "episode" OR "interview"
• "{{name}}" "{{company}}" site:youtube.com

Extract:

• Conference names and dates
• Talk titles and topics
• Podcast names and episode details
• YouTube videos (talks, interviews, webinars)
• Key quotes or positions taken

Phase 3: Written Content

WebSearch queries:

• "{{name}}" "{{company}}" "blog" OR "article" OR "wrote" OR "author"
• "{{name}}" site:medium.com
• "{{name}}" "{{company}}" site:substack.com

Extract:

• Blog posts (personal or company blog)
• Medium articles
• Substack newsletters
• Guest posts on industry publications
• Topics and themes of their writing

Phase 4: Twitter/X Activity

WebSearch queries:

• "{{name}}" "{{company}}" site:twitter.com OR site:x.com
• from:@possible_handle "{{company}}" (if handle can be inferred)

Extract:

• Twitter/X handle
• Recent tweets (topics, opinions, retweets)
• Engagement style (thought leader, curator, responder, lurker)
• Followers count (if visible)
• Notable threads or viral tweets

Phase 5: Career Trajectory

WebSearch queries:

• "{{name}}" "joins" OR "appointed" OR "promoted" OR "new role" OR "announces"
• "{{name}}" "{{company}}" "previously at" OR "former" OR "ex-"
• "{{name}}" site:crunchbase.com OR site:angel.co

Extract:

• Current tenure at company (how long?)
• Previous companies and roles
• Career progression pattern (agency to brand, brand to brand, startup founder, etc.)
• Recent job change (within last 6 months = high relevance signal)
• Board positions or advisory roles
• Investments or startup involvement

Phase 6: Shared Interests and Connections

WebSearch queries:

• "{{name}}" "{{company}}" "award" OR "recognition" OR "achievement"
• "{{name}}" "{{agency_founder_name}}" OR "{{agency_name}}" -- check for existing connections
• "{{name}}" "{{company}}" hobby OR passion OR volunteer OR community

Extract:

• Mutual connections (if any with the agency team)
• Shared alma mater, city, or industry events
• Personal interests visible in public profiles (sports, causes, hobbies)
• Awards or recognitions
• Community involvement

Phase 7: Synthesize Personalization Hooks

From all research, generate 3-5 specific, actionable personalization hooks. Each hook should be:

1. Specific: Reference a real post, talk, or event, not a generic trait
2. Recent: Prefer hooks from the last 3 months
3. Relevant: Connect to the agency's services where possible
4. Natural: Sound like something a human would notice and mention
5. Non-creepy: Avoid referencing personal/family details, locations, or anything that feels invasive

Good hooks:

• "Your LinkedIn post about the challenges of scaling a D2C brand resonated, especially the point about checkout friction."
• "Saw your talk at ShopifyConnect about mobile commerce, we've been working on exactly that with our clients."
• "Congrats on the move to BrandX, exciting time to be building their ecommerce presence."

Bad hooks:

• "I saw you went to Stanford." (too generic, feels stalkerish)
• "I noticed you live in Brooklyn." (personal, irrelevant)
• "You seem really passionate about ecommerce." (vague, could apply to anyone)

Phase 8: Output

Return structured JSON:

{
  "name": "Jane Doe",
  "title": "Head of Ecommerce",
  "company": "BrandX",
  "linkedin_url": "https://linkedin.com/in/janedoe",
  "twitter_handle": "@janedoe",
  "recent_posts": [
    {
      "platform": "LinkedIn",
      "topic": "Mobile conversion optimization for D2C brands",
      "date": "2024-01-10",
      "key_point": "Argued that most D2C brands lose 40% of mobile shoppers at checkout",
      "url": "https://linkedin.com/posts/..."
    },
    {
      "platform": "LinkedIn",
      "topic": "The role of UGC in building brand trust",
      "date": "2023-12-20",
      "key_point": "Shared data showing UGC increases conversion 2.4x vs brand content",
      "url": "https://linkedin.com/posts/..."
    }
  ],
  "talks_or_appearances": [
    {
      "event": "D2C Summit 2023",
      "topic": "Building a conversion-first product page",
      "date": "2023-11-15",
      "url": "https://youtube.com/..."
    }
  ],
  "articles": [
    {
      "title": "Why Your Shopify Store Needs a CRO Audit",
      "publication": "Medium",
      "date": "2023-10-05",
      "url": "https://medium.com/..."
    }
  ],
  "career_notes": "Joined BrandX 8 months ago from CompetitorY where she was Senior Marketing Manager. Career trajectory: agency (3 years) -> brand-side marketing (4 years) -> ecommerce leadership.",
  "interests": ["Mobile commerce", "UGC marketing", "Sustainable packaging", "Women in tech"],
  "mutual_connections": [],
  "personalization_hooks": [
    "Reference her LinkedIn post about mobile checkout friction -- directly relevant to CRO services",
    "Mention her D2C Summit talk about conversion-first product pages -- show you've done homework",
    "She joined BrandX 8 months ago -- likely still building her stack and open to agency partners",
    "Her Medium article about CRO audits makes her a warm lead -- she already believes in the value",
    "Connect over the UGC conversation -- share a relevant case study about UGC impact on conversion"
  ],
  "outreach_tone_recommendation": "Peer-to-peer, reference shared expertise in CRO. She's knowledgeable, so lead with specifics, not basics.",
  "researched_at": "2024-01-15T14:30:00Z",
  "depth": "standard",
  "search_count": 10
}

Phase 9: Confidence Assessment

Rate the research quality:

• HIGH confidence: Found LinkedIn profile, recent posts, career data, and multiple personalization hooks
• MEDIUM confidence: Found profile and some activity, but limited recent posts or content
• LOW confidence: Common name, ambiguous results, minimal public presence

If LOW confidence, note which searches were ambiguous and suggest the user verify the LinkedIn URL directly.

Example Usage

Trigger phrases:

• "Research this person before I reach out"
• "Find info on [name] at [company]"
• "Person intel for [name]"
• "What can you find about [name] for personalization?"
• "Prep outreach research on [name]"

User: Research Jane Doe, Head of Ecommerce at BrandX
Assistant: [runs 8-12 WebSearches across LinkedIn, Twitter, conferences, articles, returns structured JSON with personalization hooks]

User: Quick lookup on this LinkedIn profile: linkedin.com/in/janedoe
Assistant: [runs 3-5 focused searches using the profile as anchor, returns condensed research]