Turn expertise into
qualified attention.

Title formulas

• [System] Under Pressure: How [Audience] Handles [Threat]
• I Built [Artifact] to Test [High-Stakes Claim]
• The [Risk] Most [Role] Miss — Live Architecture Review
• From [Signal] to [Decision]: A Live [Domain] Workflow
• [Number] Design Rules for [Outcome] Without [Failure]

Search + retention description

“Watch a live, evidence-led breakdown of [problem] for [audience]. By the end, you will have [artifact/outcome]. We will examine [3 agenda points], then take technical questions. Sources and architecture notes: [link]. This is educational analysis—not legal, security, or incident-response advice. Subscribe for weekly systems briefings.”

Brand dilution risk: “shocking,” “insane,” guaranteed outcomes, or keyword stuffing.

Category decision

1. Use Science & Technology for AI agents, architecture, cybersecurity, and systems demos.
2. Use Education only when the format is explicitly tutorial-led.
3. Set language, captions, recording date, and audience correctly.
4. Add 3–5 topic phrases naturally; tags are secondary metadata, not strategy.

Discovery cluster

Anchor every live to one stable pillar: Agent Architecture, Cyber Defense, OSINT Tradecraft, or Executive Systems. Link the matching playlist, previous field brief, and next scheduled event.

Brand dilution risk: switching categories to chase a trend or attaching unrelated high-volume keywords.

Composition rules

• One idea, one focal subject, 2–4 words
• Dark technical field + glass-green accent
• Face or artifact occupies 35–45% of frame
• Readable at 10% scale; 16:9 safe margins
• Title and thumbnail create complementary gaps

Three visual concepts

Brand dilution risk: reaction faces, hazard overload, fake dashboards, or unreadable cyberpunk decoration.

Extraction system

1. Mark moments live: contrarian claim, diagram reveal, failure diagnosis, executive implication, answer.
2. Cut 20–45 seconds; show the payoff in the first 1.5 seconds.
3. Use burned-in captions, restrained motion, and a native 9:16 reframing.
4. End with a specific bridge: “Full threat model in the live briefing.”
5. Publish 3–5 clips over seven days, not all at once.

Hook bank

• “Your AI agent is not autonomous. It is under-governed.”
• “This one trust boundary changes the entire architecture.”
• “The dashboard says secure. The evidence says otherwise.”
• “Here is what an executive should ask before approving this system.”

Cadence: T+2h insight; T+1d diagram; T+3d contrarian cut; T+5d Q&A; T+7d next-live bridge.

Destination priority

1. YouTube: canonical room, replay, chapters, search equity.
2. LinkedIn: executive framing, native event, post-live document carousel.
3. X: real-time technical commentary, clips, expert conversation.
4. Optional Twitch: only for long builds with an established community.

Sequence

T−72h: canonical event → T−48h: LinkedIn thesis → T−24h: X proof thread → T−60m: native reminders → Live: monitor each chat → T+2h: replay/chapter update → T+1d: platform-native insight.

Use a destination-specific intro and CTA. Assign a moderator if chat cannot be monitored everywhere.

Brand dilution risk: unattended simulcast chats, identical captions, or sending viewers through multiple redirect hops.