Daniel Lyons' Notes

AI Assisted Writing

This workflow integrates generative AI not just for drafting, but for high-level structural planning, predictive analysis, and multimodal production.

Phase 1: Pre-Writing & Discovery (Idea Generation)

  • 1.1 Brainstorming & Idea Capture
    • Voice-to-Text Capture: Transcribe spoken ideas immediately to bypass the "blank page" friction.
    • Conversational Sparring: Use voice-enabled AI to "argue" and stress-test core concepts and premises before writing them down.
    • Competitive Intelligence Agents:
      • Gap Analysis: Use AI to scan competitor sitemaps and top-performing content to identify topics they haven't covered.
      • Trend Identification: Analyze patterns in competitor headlines and publishing frequency.
    • Zero-Shot Idea Expansion: Use "Zero-Shot" prompts (asking without examples) to generate wild, adjacent, or contrasting viewpoints based on a single seed idea to break echo chambers.
    • Smart Note-Taking:
      • "Read" Long-Form Input: Feed PDFs, YouTube URLs, or podcast transcripts into AI.
      • Timestamped Extraction: Ask for "Key insights with timestamp citations" to create a navigable database of sources.
  • 1.2 Research & Knowledge Synthesis
    • Deep Research Tools: Summarize technical documents (PDFs, whitepapers) and extract key claims/data tables.
    • Literature Gap Analysis: Feed the abstracts of 5-10 related articles and ask: "What perspective is missing from this collection?"
    • Fact Verification: Perform initial checks on complex claims and citation existence.
  • 1.3 Identify Evaluation Metrics
    • Predictive Performance Modeling:
      • Headline Scoring: Ask AI to score 10 headline variations based on "curiosity gap" and "promise of value."
      • Viral Potential Analysis: Compare your topic/angle against current trending themes to predict audience resonance (KPIs: CTR potential, Shareability).

Phase 2: Structure & Drafting (Creation)

  • 2.1 Outline Development
    • Chain-of-Thought (CoT) Logic Extraction: Instead of asking for an outline, ask the AI to "Think step-by-step to determine the most logical order of arguments for this topic," then generate the outline.
    • Reverse Outlining: Feed AI raw brain dumps and ask it to extract a structured hierarchy (Thesis Supporting Point Evidence).
    • Structural Alignment: Map the outline to known successful frameworks (e.g., PAS [Problem-Agitation-Solution], AIDA [Attention-Interest-Desire-Action], or The Hero's Journey).
  • 2.2 Draft Generation
    • Multi-Draft Generation: Generate candidates with distinct tones (e.g., "Draft A: Authoritative & Data-Heavy" vs. "Draft B: Narrative & Personal").
    • Few-Shot Style Enforcement:
      • Style Extraction: Feed the AI 3 examples of your best previous work to "teach" it your specific sentence cadence and vocabulary.
      • Sentence-Level Alignment: Apply this extracted style guide to new drafts to ensure consistency.

Phase 3: Revision & Critique (Developmental Editing)

  • 3.1 Structural Analysis
    • Pacing & Density Mapping: Ask AI to "Highlight sections that are low in information density" to spot fluff.
    • Logic Stress-Testing: Prompt the AI to act as a "Hostile Debater" to find logical fallacies, weak evidence, or plot holes.
  • 3.2 Tone & Bias Check
    • Tone Policing: Scan for passive voice or "hedging" language (e.g., "I think," "maybe") and rewrite for authority.
    • Sensitivity & Bias Reading: Flag potentially biased language, stereotypes, or cultural blind spots.
  • 3.3 Simulated Feedback
    • Persona Simulation: "Read this draft as a [Target Audience Member, e.g., Skeptical CTO] and list your top 3 objections."

Phase 4: Final Polish & Production

  • 4.1 Iteration & Proofreading
    • Standard Fixes: Grammar, spelling, and punctuation.
    • Clarity & Conciseness: "Identify sentences over 25 words and split them for better readability."
  • 4.2 Visual & Multimodal Asset Generation
    • Data Visualization: Convert text-based statistics into code for charts (Mermaid.js, Python Matplotlib, or Excel tables).
    • Image Generation: Generate prompts for cover art/thumbnails based on the draft's core metaphors.

Phase 5: Promotion & Repurposing

  • 5.1 Core Promotional Copy
    • Context-Aware Copy: Generate specific metadata: Email Subject Lines (high open rate focus), SEO Meta Descriptions, and Social Teasers.
    • SEO Optimization: Analyze the final draft against target keywords and suggest semantic additions.
  • 5.2 Format Shifting (Content Repurposing)
    • Article Thread: Extract the "top 5 punchiest insights" for X/LinkedIn threads.
    • Text Script: Convert the article into a 60-second vertical video script (Hook Value CTA).
    • Draft Audio: Generate a synthetic podcast or audio summary for accessibility and on-the-go consumption.
AI Assisted Writing
Interactive graph
On this page
Phase 1: Pre-Writing & Discovery (Idea Generation)
Phase 2: Structure & Drafting (Creation)
Phase 3: Revision & Critique (Developmental Editing)
Phase 4: Final Polish & Production
Phase 5: Promotion & Repurposing