thought-leadership

name	thought-leadership
version	1.0
last_updated	"2026-01-16T00:00:00.000Z"
author	marketing-team
description	Creates long-form thought leadership articles and whitepapers through a 3-phase human-in-the-loop process (outline, draft, polish). Produces publication-ready articles with contrarian angles, evidence-backed arguments, and TOV-matched prose. Depends on expert-pov and transcript-analysis for source material. Triggers: "thought leadership article", "long-form content", "whitepaper", "write an article about X", "expand this insight into a piece". Feeds into newsletter editions and linkedin-expert-posts as anchor content. NOT for short-form social posts — use linkedin-content instead.
goal	Creates long-form thought leadership articles and whitepapers through a 3-phase human-in-the-loop process (outline, draft, polish).
outcome	Creates long-form thought leadership articles and whitepapers through a 3-phase human-in-the-loop process (outline, draft, polish). Produces publication-ready articles with contrarian angles, evidence-backed arguments, and TOV-matched prose. Depends on expert-pov and transcript-analysis for...
primitive	content
sub_primitive	execution
ontology_type	thought-leadership
review_gate	2
inputs	{"required":["expert-pov","content-strategy"],"recommended":[]}
outputs	[{"type":"thought-leadership","feeds_into":[]}]
depends_on	["expert-pov","content-strategy"]
feeds_into	[]
owned_by_agent	content
mcps_used	["exa"]
push_targets	["gdrive","notion"]
triggers	{"slash_commands":["/thought-leadership"],"natural_language":[]}
status	draft
locked_by	null
locked_date	null
lock_version	null
sources_count	0
effort	high

Thought Leadership

When to run

Long-form articles, essays, whitepapers, or guest posts where the user has a contrarian or non-obvious angle they want to publish. Three human-approval gates (Phase 1 pyramid, Phase 2 prose, Phase 3 polish). Skip for LinkedIn posts (linkedin-content), newsletter copy, or website copy. Triggers, anti-triggers, and pre-flight checklist → references/phase-templates-and-protocols.md.

Inputs

Input	Status	Source
Topic / core insight	Required	User — must be specific and contrarian
Target audience	Required	ICP research or user-specified
Core argument	Required	User or collaboratively developed
ICP research	Recommended	clients/[client]/icp-behavioural.md
Product messaging	Recommended	clients/[client]/product-messaging.md
Content strategy	Recommended	Pillars + roadmap
Competitor positions	Recommended	Identifies white space
Source material, word count, publication context	Optional	LinkedIn 800-1500 / blog 1500-2500 / guest 2000-3500

Client CLAUDE.md is auto-loaded — apply its voice/messaging anchors. Full input matrix and pre-flight checklist in references/phase-templates-and-protocols.md.

Steps

1. Validate inputs — topic specific, audience known, contrarian angle present, user has time for 3 review cycles.
2. Phase 1.1 — Clarify core insight. Ask: "What's the one thing you want the reader to believe after reading this that they didn't believe before?"
3. Phase 1.2 — Identify audience starting point. Current belief, gap to governing thought, anticipated objections.
4. Phase 1.3 — Build the pyramid. Governing Thought → 2-4 Key Line Arguments → Supporting Evidence per key line. Use the template in references/phase-templates-and-protocols.md. Pyramid theory + MECE deep-dive in references/pyramid-principle-guide.md.
5. Phase 1.4 — Test structure. MECE check (no overlap, fully supports governing thought) + "so what?" test on every level.
6. Phase 1 deliverable + STOP. Wrap in the Phase 1 checkpoint format. Wait for explicit "Approved — proceed to Phase 2" before continuing.
7. Phase 2.1 — Map pyramid to prose sections. Governing thought → opening, key lines → sections, objections → inline or dedicated, "so what?" → closing.
8. Phase 2.2 — Write opening. Hook → Context → Thesis → optional Roadmap. Phase 2.3 — Expand each section: Headline → Lead → Evidence → Transition. Phase 2.4 — Write closing: Restate → Synthesize → Implications → CTA.
9. Phase 2 deliverable + STOP. Three title options + full prose outline. Wrap in Phase 2 checkpoint format. Wait for "Approved — proceed to Phase 3".
10. Phase 3.1 — Apply TOV. Operator authority, prescriptive confidence, framework-driven clarity, candid partnership, specificity over superlatives, conversational warmth. Em dashes with spaces, sentence case headlines, numerals, banned-word check. Quick checklist → references/tov-checklist.md.
11. Phase 3.2 — ICP guardrails. Would target ICP care? Right sophistication? Would they share this with their team?
12. Phase 3.3 — Anti-hallucination guardrails. No invented client names/metrics/quotes. Mark missing data as [Data point needed]. Verify proof points or flag as [Example — verify before publishing].
13. Phase 3.4 — 100 Posts Test. "If I published 100 posts like this, would the aggregate feel authentic to my voice and brand?" Plus final read-aloud check.
14. Phase 3 deliverable + final STOP. Wrap as THOUGHT LEADERSHIP FINAL — PHASE 3 with quality report and items-for-review. Wait for "Approved — ready to publish".
15. Push to chains. After approval, offer LinkedIn teaser, standalone quote pulls, email intro, or newsletter summary. Suggest /linkedin-content, /email-nurture, or update /content-strategy plan.

Format guidelines per platform → references/format-guidelines.md. Phase output wrappers, checkpoint dialogues, MECE details, Exa research substrate, and full auto-update protocol → references/phase-templates-and-protocols.md.

What good looks like

References — references/pyramid-principle-guide.md (Minto deep-dive), references/tov-checklist.md (voice quick-check), references/format-guidelines.md (per-platform word counts and structure), references/phase-templates-and-protocols.md (verbatim phase templates, checkpoints, anti-hallucination guardrails, integration map, auto-update protocol, MCP/Exa substrate).

Examples — examples/worked-example-gtm-engineer.md (full 3-phase walkthrough), examples/top-performing-examples.md (validated patterns).

Evaluations — Governing thought is contrarian and specific (not commodity wisdom). Key lines are MECE. Every claim has verified source or [Example — verify] flag. No banned words: solutions, leverage, synergy, scalable, best-in-class, game-changer, innovative, cutting-edge, very, really, basically, just. Three checkpoints honoured (no skipping ahead). Passes 100 Posts Test. Cites [VERIFIED: exa_search, {url}, accessed {YYYY-MM-DD}] for any external claim per .claude/rules/ontology.md and .claude/rules/exa-protocol.md.

Push

After Phase 3 approval, route the artifact to: linkedin-content (teaser posts + standalone quotes), email-nurture (newsletter share), content-strategy (add to pillar plan). Save approved final to examples/[date]-[topic-slug].md if user signals "performed well". Integration map and feedback-pattern auto-update protocol → references/phase-templates-and-protocols.md.