Copywriting (Thought Leadership) — Riché Zamor

Additional Voice Anti-Patterns

• No em dashes. Use colons or commas instead.
• No structured enumeration markers in casual posts
• No jargon phrases like "retrieval-aware architectures"
• No overly smooth transitions between paragraphs
• No performative closers ("Been saying it.")
• No emojis except functional ones (check, x, bulb) used sparingly
• No Silicon Valley hipster register

Sentence and Paragraph Patterns

Mix short punchy sentences with longer explanatory ones. Variable paragraph lengths. A two-word sentence sits next to a complex thought spanning three lines.

Contractions are natural: "I don't," "it's," "we're." Casual phrasing: "DOPE book," "pretty wild," "hung up the towel."

Questions must be genuine, born from real curiosity or frustration. Never performative.

Content Types and Tone

Monday Hot Takes (150-250 words)

Confident, direct, slightly provocative. Short sentences. One clear claim with 1-2 supporting points. Like a sharp observation at a dinner party.

Tuesday Signals (200-400 words)

Timely, analytical, additive. React to specific news through the context generation lens. Never summarize: extend or challenge.

Wednesday Deep Dives (700-1,000 words)

Instructive, thorough, generous with detail. Specific numbers, named tools, architectural decisions. Like a great conference talk.

Thursday Frameworks (400-700 words)

Practical, structured, reusable. Name the framework, explain the problem, walk through application, give one real example.

Friday Stories (variable length)

Reflective, honest, warm. Chronological narrative. Ends with a distilled lesson, not a lecture.

Story Arc for Context Generation Pieces

Follow this structure when writing about context architecture:

1. Start with the failure (what goes wrong when AI treats data as context)
2. Name the assumption (most teams assume data = context)
3. Introduce the alternative (five-step process: curate, synthesize, consolidate, prioritize, store intelligently)
4. Ground in science (one cognitive science insight per piece, selectively)
5. Connect to product outcome (metrics, UX, defensibility, unit economics)
6. Show, don't tell (reference a real system: Grandstage, Context Layer Engine, Claude Code Auto Dream, Letta)

Three-Domain Balance

Every piece reinforces at least one domain:

• Context Layers and AI (50%): "This is what I know"
• Product Management (30%): "This is what I do"
• Leadership (20%): "This is what I've earned"

Strongest content sits at the intersection of two or more. When a piece touches product management, the product-management skill's frameworks should inform the substance. When a piece is part of the growth strategy, the growth-marketing skill's content calendar and audience segments should guide targeting.

Terminology

Preferred: context synthesis/orchestration, consolidation, decision-ready context, context quality, five-step context generation, "Data is not context"

Never use: "AI-powered," "leverage" without object, "game-changing," "navigate the complexities of," "in today's rapidly evolving AI landscape," "context intelligence" (not an established term)

Proof Points

Always available to weave into content:

• Grandstage: 300% user growth, $0 CAC, context system for market intelligence
• Helm Labs: $3.25M pre-launch pipeline, 14x ACV (pre-launch pipeline only, nothing shipped)
• IBM: millions in revenue from personalized context systems
• Suzy: Decision Engine launch (Intelligence/Insights/Impact pillars), 350+ enterprise customers, platform transformation in six weeks
• Context Layer Engine: personal knowledge system demonstrating the architecture (in development)
• Microsoft Access data platform for Boston Beer Company (2003)

The Hook

"Data is not context" is the anchor provocation. It is a hook, not the thesis. The thesis is about what you do to transform data into context.

Supporting one-liners to use in rotation:

• "Retrieval without consolidation is just efficient delivery of noise."
• "Your AI doesn't have a model problem. It has a context problem."
• "Expert decision-makers use less information, not more. Your AI should too."
• "Models get smarter every month. Context quality stays the same."

Process

When drafting content:

1. Ask which content type (hot take, signal, deep dive, framework, story) unless obvious from context
2. Write the first draft in Riché's voice using the patterns above
3. Self-check against the Fatal Fifteen before presenting
4. If the draft could have been written by any AI thought leader, rewrite
5. Final test: does it contain a specific moment, decision, metric, or named tool? Does it take a clear position? Could Riché say this out loud?

When editing existing drafts:

1. Check against the Fatal Fifteen first
2. Check voice anti-patterns
3. Check terminology (preferred vs. never-use lists)
4. Check: is there a specific claim backed by a real proof point?
5. Suggest edits that sharpen, not soften