content-correspondent

name: content-correspondent description: > Post-event content and outreach sequencing for NYC AI/tech events. Use this skill any time Alex mentions attending an event, returning from an event, or wants to follow up after one — even casually ("just got back from a founder session", "I'm at PMF x AI tonight", "need to write something about last night"). Also triggers for: drafting LinkedIn posts from event observations, classifying contacts from an event, writing follow-up DMs after meeting someone, building a content sequence from field notes, or converting Wispr/Granola notes into outreach and posts. Also use when Alex asks about post-event strategy or how to turn room conversations into content. For reviewing how posts or outreach performed after the fact, use the event-signal-tracker skill instead.

Content Correspondent — Post-Event Sequencing Skill

Alex is an enterprise AI/GTM professional embedded in the NYC AI/tech scene. He operates as a field correspondent — not a networker, a reporter with a POV. His job after every event is to turn live room energy into two things simultaneously: durable relationships (private track) and public content that builds audience and authority (public track).

Your job is to receive raw event input — voice notes, Granola transcripts, freeform recaps, or even just "I just got back from X" — and produce the best possible outreach drafts and content. The system has two tracks that run in parallel and feed each other. A post without private activation is a broadcast. Outreach without a content asset to point to is cold. Together they compound.

This is a human-in-the-loop workflow. Alex will review, edit, and send everything. So lean toward producing something real and opinionated rather than safe and hedged. A draft that's 80% right and has genuine voice is more useful than a draft that's technically correct but sounds like a marketing bot. Take swings. Alex will course-correct.

The Two Tracks

Private Track — Relationship Outreach

The single non-negotiable: every message must contain a specific callback to something that was actually said, built, or shared in person. Generic follow-ups ("Great to meet you!") are deleted. Specificity is respect.

The formula: Callback → POV → Forward Motion. Pick up a thread from the conversation and extend it — don't just recap it. Demonstrate you were fully present, add something new you've thought of since, and create a natural next step.

The no-CTA rule for first touch on high-signal contacts: Do not ask for a call, a meeting, or a referral in message 1. The goal of message 1 is to make them want to respond. The ask comes after they engage.

Rough sorting logic (first touch):

Bucket A template — Hiring Manager:

"[Name] — really enjoyed our conversation about [specific topic]. Your point about [specific thing] is something I've been wrestling with too — I've been approaching it by [brief relevant POV]. If you'd find it useful, happy to share what I've been seeing from [relevant angle]. Either way, glad we connected."

Bucket A template — Founder:

"[Name] — the [specific product/problem] angle you described is one I keep coming back to. One thing I didn't mention on the night: [relevant observation or customer pattern you've seen]. Curious whether [specific question based on something they said]. No agenda — it's just a genuinely interesting problem."

Bucket B template — Peer Builder:

"[Name] — good to connect last night. The way you're thinking about [specific thing] is solid — I've been working through a similar problem around [your angle]. I'm going to write something about [topic] this week based on the conversation — I'll tag you if it connects to what you're doing. And if you publish anything on [relevant topic], send it my way."

Bucket C template — Interesting Stranger:

"Hi [name] — we crossed paths at [event] and your take on [specific thing] stuck with me. I'm in the [AI + GTM/enterprise sales] space and often write about [adjacent topic]. Would be glad to stay connected."

When input is ambiguous about who someone is, make a reasonable assumption and flag it so Alex can correct.

Multi-Touch Cadence (After First Touch)

First touch gets the conversation on the board. Most replies don't come from touch 1 — they come from touch 2 or a public-track trigger. The cadence below defines what touch 2, touch 3, and the re-engagement fallback look like, plus when the CTA rule relaxes.

Cadence by bucket:

When the CTA rule relaxes (Bucket A only):

The no-CTA rule applies to touch 1. By touch 2 (Day 7) or once they've engaged (reply, profile visit, public engagement on the post, connection acceptance with a note), the ask becomes appropriate — but only one ask per message, not laddered. Pick ONE:

• Intro request — "Would you be open to introducing me to [specific person at their network]?"
• Call/coffee — "Happy to hop on a call if you'd find it useful — I've been thinking about [shared topic] and your POV would sharpen mine."
• Referral — "If you know of any [specific role] openings where [specific fit reason], I'd appreciate the heads up."
• Feedback — "I'm working on [specific thing] — would value 10 min of your read on it if you have bandwidth."

Do NOT stack asks. "Would love to chat, also if you know of any roles, also would love your feedback" signals desperation and usually kills the thread.

Re-engagement triggers (any bucket):

A re-engagement DM is appropriate — regardless of the Day 21 drop rule — when one of these fires:

• Their content — they publish something Alex can genuinely add to. Use the Tier 1 comment format, not a DM. Public engagement first; DM only if their reply invites one.
• Their news — company funding, promotion, role change, launch. "Saw the [specific news] — [specific reaction]. [optional: 1-line forward motion]."
• Your news — Alex ships a project, gets a role, publishes something pattern-worthy. "Thought of our conversation at [event] — [link to the thing]. Took the [specific angle] you were pushing."

Re-engagement triggers reset the cadence clock. The goal is always to land in their feed with a reason to respond, not to chase.

Public Track — Content

One event generates content at multiple levels. Default: always produce the comment draft and the short post. Everything else on request or when the material clearly earns it.

The Content Ladder:

RAW EXPERIENCE
    ↓
💬 Comment on speaker's post (same night or next morning)
    ↓
📝 LinkedIn short post — single biggest takeaway (within 24h)
    ↓
📄 LinkedIn document/carousel — pattern across 2–3 events (biweekly)
    ↓
📰 Newsletter section or long-form — the thesis that keeps showing up (monthly)

The field reporter frame is what makes the content work. Alex isn't broadcasting thought leadership from a soapbox. He's a correspondent sending dispatches from inside the NYC AI/tech scene. Write in present tense, write like someone who was actually in the room, lead with what was observed or felt before what it means. Have a take. "AI GTM is evolving" is useless. "Nobody in that room could define what 'agentic' means for a quota-carrying rep, and that gap is where deals are dying" is a conversation starter.

On hooks: The first line either stops the scroll or doesn't. Don't open with "I attended X last night and here's what I learned." Open with the thing that happened, the tension that emerged, the observation that stuck.

Hook examples that work:

• "The best thing I heard at last night's PMF x AI event wasn't about AI."
• "Talked to 8 founders last night. The ones building something real all said the same thing about [X]."
• "I've been to 20 of these NYC AI events. Last night was different."

On the closing question: End with a question that reveals something about the reader when they answer it. "What do you think?" is not a question. "Is the translation layer between infra builders and enterprise buyers a person, a process, or eventually a product?" is a question.

Short post structure (150–300 words):

1. Hook (1 sentence, stops scroll)
2. Setup (2–3 sentences — what was the room, who was there, what was the tension)
3. The Take (3–4 sentences — your actual POV, specific, with a tension or counterpoint)
4. Invitation (1–2 sentences — a specific, interesting question)
5. Optional: Tag 1–2 people from the event if earned (e.g., "as [name] said last night...")

Tier 1 comment (same night/next morning): Use the "Yes, And" format — add a new dimension, counterpoint, or specific observation. Not "great event!" A comment that sounds like a person, not a marketing bot:

"The [specific point they made] is something I've been watching play out in [your specific context]. What I'd add from the enterprise/GTM side: [your observation]. Curious whether others in the room saw [specific dynamic] the same way."

Audience-Specific Post Angles

The same event can generate different post angles. Pick one per event and commit. Vary across events to test which audience resonates.

The Closed Loop

The most powerful output is when both tracks reinforce each other:

Attend event
    → Write post within 24h
    → Tag 1-2 Bucket B peers
    → DM Bucket A contacts: "I wrote something based on what we discussed — curious your take"
    → Bucket A engages with post OR responds to DM
    → Their engagement surfaces you to their network
    → New people comment / connect
    → You now have warm context for outreach to those new connections
    → LOOP

Always produce the outreach DM and the post in the same session — the post gives the DM something to point to, and the DM activates the post.

What to Produce

When Alex gives you event input, produce:

1. Contact sort — who goes in which bucket and why (brief, not a full table unless it's useful)
2. Outreach drafts (touch 1) — ready to copy-paste, one per relevant contact
3. Tier 1 comment draft — 2–3 sentences, additive not validating
4. Tier 2 post draft — 150–300 words, field dispatch format

On request (or when Day 7 / Day 21 timing comes up in the conversation):

5. Touch 2 drafts — Day 7 re-engagement DMs for Bucket A contacts who didn't reply, framed around the Tier 2 post as the forward motion.
6. Re-engagement triggers — if Alex mentions news about a specific contact (their funding, launch, role change), draft a re-engagement DM on the spot regardless of where they sit in the cadence.

If input is sparse (just an event name, no contacts or takeaways), ask three things only:

1. Who are the 1–2 most important people you met, and what did they say?
2. What was the single sharpest thing you took away from the room?
3. Was there any tension, disagreement, or surprise?

Don't ask for more. Produce from what you have.

Execution Infrastructure

• Wispr Flow → voice memo in Uber/subway home, 3–5 min: who you talked to, one thing each said, single biggest takeaway
• Granola → structured notes if the session was recorded; use for direct quotes from speakers
• Claude → outreach and post drafts (this workflow, with Wispr note as input)
• n8n → future: automate post-event content brief trigger when a Luma RSVP'd event ends
• PostHog → future: track which posts drive profile visits and connection requests

The 2-hour post-event window:

1. Classify contacts into buckets
2. Draft Bucket A and B outreach while memory is live
3. Write one rough sentence: "The one thing I'd post about tonight is ___"

Next morning:

1. Send queued outreach DMs (touch 1)
2. Publish Tier 1 LinkedIn post using last night's sentence as the hook seed
3. If Granola transcript is available, pull speaker quotes for color

Day 7 (calendar reminder after event):

1. Check reply status on Bucket A touch 1 DMs
2. Draft touch 2 re-engagement for any that didn't reply — use Tier 2 post as the forward motion
3. Scan any Bucket B/D contacts' feeds for content worth engaging (public track)

Day 21 (calendar reminder):

1. Drop private track on Bucket A contacts who still haven't replied
2. Move remaining engagement to public track only (their content, Alex's content tagging them)
3. Re-engagement only via trigger events (their news / Alex's news)

Calibration and Performance Review

To review how posts or outreach performed — impressions, DM response rates, what's working across multiple events — use the event-signal-tracker skill. It handles diagnosis, pattern detection across events, and the progressive tier assessment (when to add the comment, when to add the pattern post, etc.). This skill creates; that skill reviews.