Outreach Packager

Parse into: [Name, Company, (optional: Role, Context)]

Company-only list (no names): work at company level and note: "Add contact names for more personalized messages."

Batch limit: Up to 5 at a time. Longer lists process in batches of 5 with results between batches.

Step 3 — Research Each Prospect

For each prospect, run 2-3 web searches:

Search 1 — Company context: "[Company] [current year]" — news, funding, launches, hiring

Search 2 — Personal context (if name provided): "[Person] [Company]" — LinkedIn activity, talks, blog posts, quotes

Search 3 — Signal hunting: "[Company] hiring OR launching OR raised OR expanding OR partnering" — timely trigger

Search 4 — Warm path detection: "[Company] [your company/industry]" — shared communities, events, mutual connections

Extract:

• Company snapshot: What they do, stage, recent news (2-3 bullets)
• The signal: One specific, timely event that makes outreach relevant NOW
• Contact context: Role, background, recent public activity
• Warm path: Shared communities, events, mutual connections, content overlap
• Fit score: If ICP exists, score 1-10 on fit/timing/access/intent

No signal found? Don't fake it. Use the ICP doc + industry trends to craft a "pattern interrupt" — a relevant industry insight that demonstrates you understand their world. Flag it: "No public signal found. Using industry-angle approach."

Step 4 — Craft 3-Touch Sequence

For each prospect, write a 3-message sequence across the best channel:

Touch 1 — Day 0: The opener

• Open with the signal (or pattern interrupt if no signal)
• Connect to their likely challenge
• End with a specific question — NOT a pitch
• LinkedIn DM: 1-2 sentences max. One question. That's it.
• Email: 3-5 sentences + subject line. Short paragraphs.

Touch 2 — Day 3: The value drop

• Share something useful — a relevant article, framework, insight, or data point
• Reference your previous message without being needy
• No ask. Pure value.
• LinkedIn DM: 1-2 sentences + link/resource
• Email: 2-4 sentences + the resource

Touch 3 — Day 7: The direct ask

• Acknowledge you've reached out before
• Make a specific, time-bound ask (15-min call about X)
• Give them an easy out ("If this isn't relevant, no worries")
• LinkedIn DM: 2-3 sentences
• Email: 3-5 sentences

Channel recommendation: Default to the channel where the prospect is most active. If they post on LinkedIn regularly → LinkedIn DM. If they have a public email or blog → email. If unsure → provide both.

Also generate an A/B variant for Touch 1 — same signal, different angle (e.g., one casual, one more direct).

Step 5 — Present Results

For each prospect:

---

## [Name] — [Company]

**Company:** [What they do — 1 sentence]
**Recent signal:** [The timely event — or "No signal: industry angle"]
**Warm path:** [Shared connection/community — or "None found"]
**Why reach out now:** [1 sentence connecting signal to value]
**Fit score:** [X/10] (if ICP available)
**Recommended channel:** [LinkedIn DM / Email / Both]

### Touch 1A — Day 0 (Primary)
[Message]

### Touch 1B — Day 0 (A/B Variant)
[Message — different angle]

### Touch 2 — Day 3 (Value Drop)
[Message]

### Touch 3 — Day 7 (Direct Ask)
[Message]

---

After all prospects: "Researched [N] prospects. [X] with strong signals, [Y] with industry angles. [Z] with warm paths — prioritize those."

Step 6 — Handle Revisions

If user gives feedback on tone or approach:

• Fix immediately
• If generalizable (e.g., "I never mention competitors in first touch"), apply to all remaining messages

More prospects? "Got more names? Drop them and I'll research the next batch."

Step 7 — Feedback Loop (Reply Tracking)

This step triggers when the user says "track replies", "which outreach worked", "outreach results", or returns after sending a batch.

1. Ask: "Which messages got replies? For each, tell me: replied / no reply / meeting booked. And which touch # got the reply?"

2. Log results in docs/outreach-log.md (create if doesn't exist):

# Outreach Performance Log

## Batch [date] — [N] prospects

| Prospect | Signal Used | Channel | Touch # | Result |
|----------|-------------|---------|---------|--------|
| [Name] at [Company] | [Hiring for X] | LinkedIn DM | Touch 1 | ★ Meeting booked |
| [Name] at [Company] | [Series B] | Email | Touch 3 | ○ Replied, no meeting |
| [Name] at [Company] | [Industry angle] | LinkedIn DM | — | ✗ No reply |

**Reply rate:** [X]%
**Best signal type:** [Which signal type got replies]
**Best channel:** [DM vs email]
**Best touch:** [Which touch # got most replies]
**Warm path impact:** [Did warm-path prospects reply more?]

3. Analyze patterns across all logged batches:

   • Which signal types get replies? (hiring > funding > product launch?)
   • Which channel performs better? (DM vs email)
   • Which touch # gets the most replies? (1, 2, or 3?)
   • Do warm-path prospects convert better?
   • What message length gets responses?
   • If ICP exists, which scoring dimensions correlate with replies?

4. Apply learnings to next batch:

   • Prioritize signal types that historically convert
   • Lead with the better-performing channel
   • Tell the user: "Based on your last [N] batches, hiring signals get 3x more replies than funding news. Touch 2 value drops have a 25% reply rate. I'll optimize accordingly."

5. After 3+ batches, recommend ICP refinements: "Prospects scoring 8+ on Timing replied 40% of the time. Prospects below 5 never replied. Consider raising your Timing threshold."

Common Mistakes

• Cold DMs that are too long. LinkedIn DMs over 2 sentences don't get read. Be brutally short.
• Faking a signal. If there's no public signal, say so and use an industry angle instead. "I noticed your company is doing great things" fools nobody.
• Only sending one message. 80% of replies come from touch 2 or 3. The sequence IS the strategy.
• Same message, different name. The signal and angle should be genuinely different per prospect. If all messages look the same, you're spraying.
• Pitching in Touch 1. First touch = earn a conversation, not close a deal.
• Processing too many at once. Quality drops past 5. Batch strictly.
• Never tracking what got replies. Sending outreach without logging results means you never learn which signals convert. Track every batch.
• Ignoring warm paths. A shared community or mutual connection is worth 10x a cold signal. Always check.

Key File Paths