khetansarvesh/linkedin-outreach

LinkedIn Outreach Message Drafter

Draft personalized LinkedIn outreach messages that maximize response rates. Takes a LinkedIn post URL, researches the person and company, then generates a concise 3-paragraph message.

When to Activate

• User shares a LinkedIn post URL and wants to reach out
• User wants to message a hiring manager, recruiter, or team lead about a role
• User asks to draft a LinkedIn connection request or InMail

Profile Data

Read the user's profile from these files before drafting:

• Notion (fetch via python3 scripts/notion/page_reader.py resume) — personal info, experience, skills
• Notion (fetch via python3 scripts/notion/page_reader.py projects) — project summaries for matching relevance

Workflow

Step 1: Research the Target

Use WebFetch on the LinkedIn post URL to extract:

• Who posted: name, title, company
• What the post is about: job opening, company news, product launch, thought leadership
• Company details: what they do, industry, stage, size
• Role details: if hiring, what skills/qualities they want

If the post doesn't have enough info, use WebSearch to research:

• The company (what they build, recent news, funding, mission)
• The person (their role, background, what they care about)

Step 1.5: Classify Contact Type

Determine who you're messaging — this changes the message framework:

| Type | Who | Message Focus | | ------------------ | ------------------------------------------- | ------------------------------------------------- | | Recruiter | Talent acquisition, sourcing | Fit criteria → screening answers → CTA | | Hiring Manager | Team lead who's hiring | Specific challenge → quantified achievement → ask | | Peer | Someone in a similar role on the team | Genuine interest → shared problem → soft ask | | Interviewer | Someone who will interview you (date known) | Research signal → connection → looking forward |

Key rule for Peer outreach: Do NOT ask for a job. The referral happens naturally if the conversation flows. Lead with genuine interest in their work.

Key rule for Interviewer outreach: Keep it light. The goal is they know you prepared, not that you're desperate.

Step 2: Find the Connection

From the user's profile, identify the single strongest overlap between:

• User's experience/projects and what the company does
• User's skills and what the role needs
• User's interests and what the person posted about

Pick ONE specific thing — not a laundry list.

Step 3: Draft the Message

Write a 3-paragraph LinkedIn message following these rules:

Structure by Contact Type

For Recruiters (3 sentences):

1. Fit: Direct match criteria — role, relevant experience, availability/location
2. Proof: Answer their screening questions before they ask ("3+ years building ML pipelines, MS in ML from UMD, available June 2026")
3. CTA: "Happy to share my CV if this aligns with what you're looking for"

For Hiring Managers (3 sentences):

1. Hook: Specific challenge their team faces (from JD, company blog, or news)
2. Proof: Your biggest quantified achievement that shows you've solved similar problems
3. CTA: "Would love to hear how your team is approaching [specific challenge]"

For Peers — referral (3 sentences):

1. Interest: Genuine reference to their work — blog post, talk, open-source project, paper
2. Connection: Something you're doing in the same space (NOT a job pitch)
3. CTA: "I've been working on similar problems at [company], would love to hear your take on [topic]"

For Interviewers — pre-interview (3 sentences):

1. Research: Reference something specific about their work or background
2. Context: Light connection to your experience in that area
3. CTA: "Looking forward to our conversation on [date]"

Generic (when type is unclear):

Sentence 1 — Hook + About Me

• Reference their specific post/role (shows homework)
• One sentence on who you are + most relevant achievement

Sentence 2 — Why I'm a Fit

• Connect ONE specific project/skill to what they need
• Use a concrete metric (not vague claims)

Sentence 3 — Soft Ask

• Low-pressure ask for a quick chat
• Easy out so it doesn't feel transactional

Rules

• Connection request: max 300 characters (LinkedIn limit for connection notes)
• InMail/message: 80-120 words max — shorter messages get higher response rates
• Plain simple English — no corporate jargon, no buzzword soup
• No flattery — don't say "I love your company" or "your post was amazing"
• No desperation — don't say "I would be honored" or "it would mean the world"
• Specific over generic — reference their actual post content, not vague praise
• One ask only — don't ask for a referral AND a chat AND feedback
• No resume dump — pick one achievement, not five
• Conversational tone — write like a human, not a cover letter
• No emojis — keep it professional but warm
• Never share phone number in the message
• Suggest alternative targets — if primary contact doesn't respond, suggest 1-2 backup contacts with justification

What NOT to write

BAD: "Hi! I came across your post and was really impressed by SmarterDx's work in healthcare AI. I'm a passionate ML engineer with 3+ years of experience in NLP, RAG, LLMs, and deep learning..."

GOOD: "Hi [Name] — saw your post about the Staff ML Engineer role at SmarterDx. I've been building production AI agents at Sentient, most recently a search agent that beat SOTA by 10% on hierarchical tasks."

Step 4: Output to Markdown File

Write all results to an output markdown file (e.g., outreach_messages.md or a name the user specifies).

The file must have:

1. A summary table at the top with all posts
2. Individual sections below with the full drafted message for each post

Output Format

# LinkedIn Outreach Messages

Generated: [date]

## Summary

The summary table must be comprehensive — it should contain ALL context (post link, location, who to message, why I'm a fit, flags) so the user can understand everything from the table alone without reading individual sections.

| #   | Company                       | Role   | Location      | Person to Message | Post        | Why I'm a Fit                                 | Flags           |
| --- | ----------------------------- | ------ | ------------- | ----------------- | ----------- | --------------------------------------------- | --------------- |
| 1   | [[Company]](#1-company--role) | [Role] | [City/Remote] | [Name, Title]     | [post](url) | [specific match — reference projects/metrics] | [concerns or —] |
| 2   | ...                           | ...    | ...           | ...               | ...         | ...                                           | ...             |

The Company column links to the message section below (anchor format: `#1-company--role` using lowercase, hyphens, no special chars).

---

## 1. [Company] — [Role]

> Hi [Name] — saw your post about the [Role] at [Company]. I'm currently building
> [specific thing] at [Company], where I [specific result with metric].
>
> [Company]'s work on [specific thing from post] is close to what I did at [previous
>
> > company] — I [specific achievement that maps to their need].
>
> Would you be open to a quick chat about the role?

---

## 2. [Company] — [Role]

...

Individual message sections contain ONLY the heading and the quoted message — no post link, match, or flags (those are already in the summary table).

If a post should be skipped (e.g., wrong location, unrelated role), still include it in the summary table with the reason in the Flags column, and write "Skipped — [reason]" in the message section.

Adapting to Post Types

| Post Type | Approach | | -------------------------- | -------------------------------------------------------------------- | | Job posting | Reference the specific role, match your skills to their requirements | | Company news/milestone | Congratulate briefly, connect your work to their growth area | | Thought leadership | Reference their specific point, share a related experience | | Product launch | Connect your experience to the problem their product solves |