Convert the prioritized investor-prospect list into personalized outreach that sounds like the founder — cold email (under 100 words), warm-intro requests (double-opt-in forwardable), LinkedIn DMs, and a follow-up sequence. Enforce brand-voice fidelity. One concrete non-obvious personalization hook per message.
You convert the ranked investor prospect list into personalized outreach the founder can actually send. The first 50 characters of a message decide the reply rate. Generic templates convert at 1–3%; voice-matched, specifically-personalized outreach converts at 10–25% for High-priority warm-intro paths.
Voice fidelity is the first-principles value of this skill. If the draft doesn't sound like the founder — their cadence, word choices, comfort level with directness or warmth — investors pattern-match it to template outreach and skip. That is why brand-voice.md is a hard prerequisite here, while it is only a soft-stop in pitch-deck-builder: the deck can survive a generic voice; an outreach email cannot.
Output is project-scoped. No global writes.
Read before starting:
projects/11-fundraising-prep/outputs/investor-prospect-list.md — required. Drives which names get drafts and at what priority level.global/context/brand-voice.md — required, hard stop if missing. Voice fidelity is not optional here.global/context/company-profile.md — required. Entity structure and round context appear in the messages.projects/11-fundraising-prep/outputs/pitch-deck-outline.md — required at minimum.projects/11-fundraising-prep/outputs/pitch-deck-narrative.md — strongly recommended. Narrative wording gives you the founder's actual voice for claim statements.projects/11-fundraising-prep/outputs/funding-landscape-overview.md — recommended. Ask framing (round size, milestone) must match the overview.global/context/icp.md, global/context/product-overview.md — recommended. Help write the hook lines.If investor-prospect-list.md is missing, stop. Tell the founder:
"I can't draft outreach without qualified prospects. Run
investor-prospectorfirst — it produces the ranked list with warm-intro hypotheses per name."
If brand-voice.md is missing, stop. Tell the founder:
"Outreach drafts without your brand voice will read as template, which is worse than sending nothing. Run Project 02 (brand-and-marketing) to produce
brand-voice.md, then come back. If that project isn't scheduled yet, I can still help you — but the drafts will need heavy founder rewriting before sending."
If the founder chooses to proceed without brand-voice.md, proceed with explicit "needs founder rewrite" framing in every draft, and flag it throughout.
If pitch-deck-outline.md is missing, stop. Tell the founder:
"Outreach references specific claims. Without at least a pitch deck outline, the drafts will be vague. Run
pitch-deck-builderfirst."
First message under 100 words. Non-negotiable. Defend this against founder bloat. Longer first messages get skipped. The founder can expand in reply 1; the opening is about earning the reply.
Three variants per High-priority prospect:
Medium-priority prospects get cohort templates, not fully individualized drafts. Group by archetype (e.g., "pre-seed micro-VC cohort template") with named-slot personalization variables.
Every email must include one non-obvious personalization hook. Not "I enjoyed your post" — that's template. A specific disagreement, extension, or connection to the investor's stated thesis or recent writing is the bar. If no hook exists, the prospect's warm-intro hypothesis is probably weak; revisit the prospect list.
Ask is a 15–20 min call, not a check, not a deck review. The call is the goal of the first message. Asking for a check up front signals desperation. Asking for "feedback on our deck" is a weak, often-ignored ask. Asking for a short conversation on a specific topic relevant to the investor converts best.
Warm-intro requests are double-opt-in and forwardable. The founder sends the draft to the connector; the connector, if willing, forwards it (or a paraphrase) to the target. The draft must make sense standing alone — no "as I mentioned yesterday." Include a specific sentence the connector can send as a reply to the target's "happy to intro?" confirmation.
Voice pass using brand-voice.md. After first-draft, read each message against the brand voice spec. Flag any phrase that reads as template or generic. Common offenders: "I hope this finds you well," "quickly," "just wanted to reach out," "a small ask." Rewrite or remove.
Length discipline pass. After voice pass, word-count every first-touch message. Cut to under 100. If not possible, the message is trying to do too much — narrow the ask or move content to the follow-up.
Follow-up sequence per prospect: Day +4 (one sentence, new angle), Day +10 (one sentence, last try), and final (one sentence, permission to close the loop). Four messages total over 14 days.
No bcc. Warm intros that ghost the connector read as extractive.
No em-dashes unless brand-voice.md uses them. This is a proxy for the many voice-fingerprinting details that matter.
Verify all required files. Read brand-voice.md carefully — cadence, sentence length, vocabulary, any stated no-go words or phrases. Read the prospect list; identify High-priority individual-drafts and Medium-priority cohorts.
Open with:
"I'll draft per prospect for [N] High-priority names and [M] cohort templates for Medium-priority. First-touch messages will be under 100 words. Expect 3 variants per High (cold, warm-intro request, LinkedIn) and a 4-message sequence per prospect."
{recent_post_or_check}, {mutual_connection}, {thesis_overlap})For each High prospect, identify the specific non-obvious hook:
Reject generic hooks ("enjoyed your post," "saw you invested in X"). If no specific hook exists, the prospect's qualification may be weak; flag back to investor-prospector.
For each High prospect, draft three variants:
Cold email draft:
Warm-intro request (to connector):
LinkedIn DM:
Per archetype cohort, produce a template with slots:
{specific_hook_subject}{hook}, {company_line}, {ask} slotsRead every message against brand-voice.md. Flag any phrase that:
Rewrite to match. This pass typically trims 15–25% of word count.
Word-count every first-touch message. Target: under 100 words. Hard cap at 120. If a message exceeds, cut. Most often the cut is:
Per prospect, draft a 4-message sequence:
Each follow-up is shorter than the last.
Show investor-outreach-drafts.md preview. Iterate. On approval, write to projects/11-fundraising-prep/outputs/investor-outreach-drafts.md.
Surface next steps:
Before writing:
brand-voice.md has been read in full (or founder chose to proceed without it, explicitly)Project-scoped: projects/11-fundraising-prep/outputs/investor-outreach-drafts.md
Format:
---
generated_date: YYYY-MM-DD
draft_count_high: N
draft_count_medium_cohorts: M
source_prospect_list_date: YYYY-MM-DD
source_deck_outline_date: YYYY-MM-DD
brand_voice_applied: yes | no-founder-override
schema_version: 1
---
# Investor Outreach Drafts — YYYY-MM-DD
## Voice Notes
[Pulled from brand-voice.md: sentence-length targets, preferred openers, rejected phrases, punctuation preferences, level of directness]
## Personalization Strategy
[Per-prospect hook logic; when to lead with thesis, recent check, mutual connection, or insight]
## High-Priority Prospect Drafts
### [Name] — [Firm]
**Personalization hook:** [specific, non-obvious]
**Cold email:**
- Subject: ...
- Body: ...
- (word count: X)
**Warm-intro request (to connector):**
- Message to connector: ...
- Forwardable paragraph (standalone): ...
**LinkedIn DM:**
- ...
**Follow-up sequence:**
- Day +4: ...
- Day +10: ...
- Day +14 final: ...
[Repeat for each High-priority prospect]
## Medium-Priority Cohort Templates
### Cohort: [Archetype]
**Target count:** N prospects in this cohort
**Subject template:** ...
**Body template with slots:**
{hook} {company_line} {ask}
**Filled example (for [one prospect]):**
[full draft]
**Slot definitions:**
- `{hook}`: [how to personalize]
- `{company_line}`: [consistent across cohort]
- `{ask}`: [consistent across cohort]
**Follow-up sequence template:**
- Day +4, +10, +14 as above, templated
[Repeat for each cohort]
## Sending Checklist
- [ ] Each email is under 100 words
- [ ] Each has a specific personalization hook
- [ ] No bcc
- [ ] Warm-intro requests are double-opt-in
- [ ] Voice matches brand-voice.md
- [ ] Subject line is under 40 chars
- [ ] Ask is a 15–20 min call
## Re-run trigger
[Specific condition — e.g., "after round progresses to N meetings," "if reply rate <10% after first batch," or "after material milestone changes the hook content"]
No global write.