Quartz 4

Workshop - AI Content Live Prompting

15 min read

Summary for 001-ai-content-live-prompting-session.mp4

Segment 1

Overarching Principles

  • Authenticity at Scale: Use AI not to generate generic content, but to scale your unique voice, expertise, and perspective. The tool should amplify your authentic “spiky point of view,” not replace it.
  • Platform Focus (LinkedIn): Concentrate content efforts on LinkedIn because it is the world’s largest professional network, where CEOs and decision-makers are active. Its algorithm prioritizes native, educational content that sparks discussion.
  • Efficiency and ROI: Leverage AI to dramatically reduce the time and cost of content creation. The process provides a clear return on investment by replacing tasks that would otherwise require manual writing or expensive ghostwriters.
  • Systematized Content Creation: The process of defining your perspective and training the AI is a reusable “Lego block” system that can be applied to create various assets, including articles, newsletters, webinar scripts, and course materials.

Frameworks

  1. Spiky Point of View (POV) Framework: To stand out, your thought leadership must be built on a unique perspective that achieves three things:

    • Rattle the Cage: Is your perspective debatable yet defensible? It should offer a fresh, genuine take on familiar situations.
    • Reveal Insights: Does it help your audience see their problem in a new, surprising way that still feels true?
    • Resonate Deeply: Does it speak to your audience’s key challenges and aspirations, making them think, “I need to hear more”?

  2. Gemini CLI / Maestro Project “Training” Framework: This method involves creating a dedicated “Project” within the Gemini CLI / Maestro AI to act as a custom knowledge base, effectively training the AI on your unique context.

    • Objective: To streamline content creation while maintaining your authentic style and substance.
    • Inputs: Gather a variety of text-based files that encompass your expertise and communication style. The key source materials are:
      • Your Biography: A comprehensive version is ideal. молекулы Transcripts: From presentations, workshops, and advisory calls.
      • Existing Text: Any articles, newsletters, emails, or documents you have previously written.

Actionable Flight Plan

  1. Uncover Your Unique Perspective:

    • Access and make a copy of the “Develop Your Unique Perspective” worksheet.
    • Dedicate 20-30 minutes to answer the reflection questions to capture the unfiltered essence of your views.
    • Distill your responses into 3-5 core “Thought Leadership Pillars” to serve as the north star for your content.

  2. Set Up Your Gemini CLI / Maestro Project:

    • Obtain a paid subscription to Gemini CLI / Maestro.ai to access the Projects feature.
    • Create a new “Project” dedicated to your personal or brand’s thought leadership.
    • Upload key documents to the Project to train the AI on your voice, style, and expertise (e.g., your bio, transcripts of talks, past articles).

  3. Generate a 30-Day Content Strategy:

    • Within your new Gemini CLI / Maestro Project, use the provided prompt frameworks.
    • Instruct Gemini CLI / Maestro to develop a 30-day content plan based on your defined Thought Leadership Pillars and the uploaded source material.

  4. Create and Refine Thought Leadership Articles:

    • Use the structured prompts to have Gemini CLI / Maestro generate drafts of LinkedIn articles and posts.
    • Engage in a collaborative “working session” with the AI to edit the drafts. Prompt it to refine tone, add analogies, or restructure sections until the final output is polished and authentic to your voice.

  5. Publish for Impact:

    • Publish the finalized content natively on LinkedIn (i.e., directly on the platform, not through a third-party scheduler) to gain algorithmic priority.
    • Focus on generating unique, educational content that is designed to spark discussion and attract your ideal clients.

Segment 2

Overarching Principles

  • Context is King: The quality and authenticity of AI-generated content are directly proportional to the quantity and quality of contextual information provided. To get the best results, you must train the AI on your unique experiences, voice, and perspective by providing it with a rich set of personal documents.
  • Authenticity Over “Professionalism”: The most effective content sounds natural and human-like. The goal is to replicate your authentic communication style, not to generate generic, formal business writing. Use transcripts of spoken content (talks, workshops) to capture your natural voice.
  • Strategic Prompting: Don’t just ask for content. Guide the AI with a structured “megaprompt” that defines your audience, goals, content pillars, unique philosophy, and desired output format to ensure the results align with your brand and strategy.

Frameworks

  1. Gemini CLI / Maestro Project Setup: A methodology for creating a centralized knowledge base for the AI by gathering and uploading a variety of text-based files. The essential components are:

    • Biography: A comprehensive account of your professional history and experiences.
    • Transcripts: Verbatim text from presentations, workshops, and advisory calls.
    • Text: A collection of written materials like articles, newsletters, emails, and internal strategy documents.

  2. Content Strategy Megaprompt: A multi-part prompt structure designed to generate a comprehensive, personalized content plan.

    • Objective & Niche: State the goal (e.g., a 30-day LinkedIn content plan) and the specific target audience.
    • Content Pillars: List 3-5 core themes you want to focus on.
    • Principles & Mental Models: Inject your unique philosophy, values, and problem-solving approaches to avoid generic output.
    • Content Cadence: Specify the desired types and frequency of content (e.g., 3 articles and 2 posts per week).
    • Output Format: Instruct the AI on how to structure each individual content idea.
    • Contextual Reference: Command the AI to reference the knowledge from the uploaded project files.

  3. Story-Insight-Takeaway (S-I-T) Format: A simple yet powerful structure for creating engaging content.

    • Story: A brief anecdote or example that illustrates the main point and hooks the reader.
    • Insight: The key lesson from the story, presented through your unique lens.
    • Takeaway: A clear, actionable way for the reader to apply the insight to their own work or life.

Actionable Flight Plan

  1. Gather Your Assets: Collect a comprehensive set of text-based files that represent your experience and voice, including your bio, transcripts of talks, past articles, newsletters, and strategic documents.
  2. Set Up Your Gemini CLI / Maestro Project:
    • Create a new Project in Gemini CLI / Maestro (this may require a paid plan).
    • Upload all the assets gathered in Step 1 to this project. This trains Gemini CLI / Maestro on your specific knowledge base.
  3. Personalize the Megaprompt:
    • Copy the “Content Strategy Megaprompt” template.
    • Fill in the bracketed, placeholder sections ([SPECIFIC NICHE], [LIST CONTENT PILLARS], etc.) with your own specific information.
  4. Generate the Content Plan: Paste the fully personalized megaprompt into your Gemini CLI / Maestro Project chat and execute it.
  5. Review and Iterate:
    • Analyze the generated 30-day content plan, including the titles and outlines.
    • Identify which ideas resonate with your voice and which do not.
    • Provide this feedback to Gemini CLI / Maestro in a follow-up prompt (e.g., “I like ideas 1, 3, and 5. I would never talk about topic 2. Generate more ideas like 1, 3, and 5.”) to further refine the output.

Segment 3

Coaching Media Analysis

Overarching Principles

  • AI as a Starting Point: AI-generated content plans are a powerful first draft, but they require human insight and personalization to be truly effective.
  • The “Spiky Point of View”: Generic or outdated concepts must be challenged with a unique, contrarian, or modern perspective to create content that stands out. The value is not just in the topic, but in your unique take on it.
  • Iterative Refinement: The most efficient way to align AI output with your brand voice is through a structured, iterative feedback loop, rather than editing one piece at a time.
  • Content Authenticity: A successful content plan reflects your actual experiences, stories, and beliefs. Generic topics should be re-framed through the lens of your personal journey and expertise.

Frameworks

The First-Pass Scan & Refinement Framework

This framework is designed to quickly and efficiently transform a generic AI-generated content plan into a personalized and authentic strategy.

  1. Initial Generation: Prompt an AI to create a baseline content plan (e.g., a 30-day plan for a specific niche).
  2. First-Pass Scan & Triage: Quickly review the entire plan to identify:
    • Strong Fits: Topics that align well with your expertise.
    • Generic/Outdated Topics: Well-worn concepts that lack a fresh angle (e.g., the “2-Minute Rule,” “Work-Life Balance”).
    • Misalignments: Topics that don’t fit your perspective or audience.
  3. Spiky POV Injection: For each generic or outdated topic, articulate your unique counter-argument or modern twist. This is the core of personalizing the content.
    • Example: The generic “2-Minute Rule” for productivity is challenged by the spiky POV that it can lead to a “death by papercut” of minor, inefficient tasks that should be automated.
  4. Consolidated Feedback Prompt: Group all your feedback, new angles, and topics you dislike into a single, comprehensive prompt for the AI.
  5. Regeneration: Instruct the AI to regenerate the entire content plan based on your consolidated feedback, resulting in a much more aligned and personalized second draft.

Actionable Flight Plan

  1. Generate a Baseline: Use an AI to generate a 30-day content strategy for your target audience (e.g., CEOs, CTOs).
  2. Conduct a “First-Pass Scan”: Read through the entire list of proposed articles and posts.
  3. Categorize and Critique: In your AI chat, create two sections: “Ones I like” and “Ones I don’t like.”
  4. Inject Your Perspective: For each topic, add specific, nuanced feedback.
    • For “Dislikes”: Clearly state why the topic is a poor fit and what you’d like to focus on instead.
      • Example: “I don’t like ‘The 2-minute rule’ - it can lead to ‘death by papercut’ and inefficiency. I want to introduce the concept of a ‘time budget’ for these small tasks to maintain focus on larger priorities.”
    • For “Generic Topics”: Acknowledge the topic is common and provide a fresh, personal take.
      • Example: “Work-life Balance is a generic topic for my audience. I need a fresh take that gives a ‘whole life’ perspective on how we use our energy in service of others.”
  5. Create a Regeneration Prompt: After providing feedback on several items, consolidate your intent with a clear instruction.
    • Example: “Can you run with these and regenerate a 30-day content strategy that aligns more closely with my expertise and audience needs?”
  6. Execute and Build: Use the revised, more personalized content plan as the new foundation to generate individual articles and posts, continuing the iterative feedback process for each piece.

Segment 4

Overarching Principles

  • Iterative Refinement over One-Shot Generation: The most effective use of AI for content creation is not a single, perfect prompt but a conversational, iterative process of providing feedback and regenerating until the output is aligned.
  • Specificity In, Specificity Out: Generic prompts yield generic content. Providing nuanced, specific feedback (e.g., changing an angle from “why micromanagement is bad” to “how to spot and redirect it”) is critical for generating valuable, non-obvious content.
  • Human Expertise as the Director: The user’s role is to be the expert, strategist, and director. The AI is the skilled but un-opinionated writer. The human provides the “what” and the “why,” and the AI executes the “how” (drafting).
  • Injecting Authenticity with Scenarios: To move beyond generic templates, feed the AI with concrete details, either through personal “brain-dump” transcripts or detailed hypothetical scenarios. This grounds the AI’s output in relatable, realistic situations.
  • Stylistic Steering: Voice and tone can be explicitly shaped by referencing stylistic examples (e.g., “more like Seth Godin”) and giving structural commands (e.g., “use short paragraphs,” “vary sentence length”).

Framework: The AI Content Co-Creation Loop

This is a structured process for moving from a high-level content idea to a polished, personalized article using an AI partner.

  1. Phase 1: Strategy Generation:

    • Input: A detailed initial prompt outlining the goal (e.g., 30-day content plan), target audience, content pillars, and desired format (e.g., Story, Insight, Takeaway).
    • Output: The AI generates a high-level content strategy or list of article ideas.

  2. Phase 2: Strategic Refinement:

    • Input: The user provides specific, directional feedback on the AI’s initial ideas, challenging generic concepts and suggesting more nuanced angles.
    • Output: The AI regenerates the content plan, aligning it more closely with the user’s expertise and audience needs.

  3. Phase 3: Context Injection & Drafting:

    • Input: The user selects a refined idea and provides the core “raw material”—either a personal story, a detailed hypothetical scenario, or a list of key talking points.
    • Output: The AI generates a first draft of the full article, structuring the user’s raw input into the desired format.

  4. Phase 4: Stylistic Polishing:

    • Input: The user provides feedback on the draft’s tone, pacing, and style, referencing other writers or specific formatting rules.
    • Output: The AI revises the draft, incorporating the stylistic adjustments to better match the user’s desired voice.

Actionable Flight Plan

  • Step 1: Define Your Strategy and Generate a Plan: Create a detailed initial prompt for your content plan. Ask the AI to generate a list of article and post ideas for a set period (e.g., 30 days).
  • Step 2: Critique and Refine the Plan: Review the AI’s suggestions. For each idea, provide feedback to make it less generic.
    • Example Prompt: “For idea #1, ‘The Hidden Cost of Micromanagement,’ the angle is too obvious. Instead, let’s focus on ‘Spotting and Redirecting Micromanagement.’ The story should be about subtly redirecting a leader who isn’t aware they’re micromanaging.”
  • Step 3: Create a Fictional Scenario: To bring a topic to life, create a detailed hypothetical situation.
    • Example Prompt: “Let’s draft article #1. I want to talk through a hypothetical leadership team meeting where the CEO declares a massive shift in strategy, bypassing weeks of work the team has done. I see the energy shift, people get dejected, and all eyes look to me. Can you draft an article with this in mind?”
  • Step 4: Incorporate Stylistic Influences: Once you have a draft, steer the AI’s writing style to match your preferences.
    • Example Prompt: “This is a good start. Now, revise it with these points in mind:
      • Make the paragraphs shorter, 20% more like Seth Godin.
      • Vary the sentence structure with very short and long sentences.
      • Make this line less corporate jargon: ‘Thank you for sharing this exciting new direction.’ It reads a little condescending.”
  • Step 5: Set a Clear Goal for the Output: Explicitly state the objective of your communication to help the AI focus its language.
    • Example Prompt: “My goal for this article is to be impossible to misunderstand. Please revise the draft to ensure the key message is clear and direct.”

Segment 5

Overarching Principles

  • Front-Loaded Effort for Exponential Returns: Invest time upfront to refine the first piece of content with an LLM. This initial “training” within a single conversation thread creates a contextual model of your preferences, drastically improving the quality and reducing the editing time for all subsequent content generated in that same thread.
  • Conversational Context is Key: Treat the LLM as a partner in a single, continuous project conversation. The AI learns from your feedback and applies that learning to future requests within the same thread, making the process progressively more efficient.
  • Develop a Personal Quality Barometer: The ultimate measure of success is not just that the content is generated, but that it meets your personal standard of quality. The speaker uses a “gut check” and asks, “Would I be proud if my most prized connection saw this?” AI output should be a tool, not a replacement for your judgment.
  • From Raw Ideas to Polished Narrative: The process effectively transforms your core ideas, experiences, and viewpoints (the “raw data”) into well-structured, story-based, and actionable thought leadership content.

Frameworks

The Single-Thread Training Framework

This framework leverages a single, continuous LLM conversation to create a series of high-quality, stylistically-consistent content pieces.

  1. Initiation & First Draft: Start a new chat thread for a specific content series (e.g., “30-Day LinkedIn Plan”). Provide the prompt for the very first article.
  2. Iterative Refinement & Feedback Loop: Critically review the first draft. Provide specific, constructive feedback to the LLM to edit and revise the article. Continue this loop until the article meets your quality standard.
  3. Context-Setting Command: Once the first article is perfected, give the LLM an explicit instruction to internalize the previous edits as a style guide for all future work in the thread.
    • Example Prompt: This is great. Thank you for the edits. Now let's start with #2 - please apply everything you've learned about my communication style and preferences for this one.
  4. Sequential Generation: Prompt for the subsequent articles in the series one by one, using simple commands (e.g., Now, let's do article #3). The AI will use the established context to produce higher-quality first drafts.

Actionable Flight Plan

  1. Define Your Content Campaign: Start with a clear objective for a series of related articles (e.g., a 10-part series on leadership, a 4-week newsletter plan).
  2. Generate Article #1: Use your initial prompt to create the first piece of content in the series.
  3. Refine Article #1 with Feedback: Review the draft and provide specific edits.
    • Example Feedback: "This is too choppy and dramatic. Please add more context, flesh out the story, and adopt a more informative, professional tone."
  4. “Lock In” the Style: Once Article #1 is finalized, use a context-setting command to instruct the AI to use it as a template.
    • Action: In the same chat, type: Excellent. Now, using the style, tone, and preferences from our edits on the previous article, please write Article #2 on the topic of [Your Next Topic].
  5. Batch-Produce the Series: Continue prompting for the remaining articles (#3, #4, etc.) in the same thread. The time required for editing should decrease with each new article.
  6. Establish a “Gut Check” Quality Gate: Before publishing any piece, review it against your personal brand standards. Ask yourself:
    • Does this sound like me?
    • Am I excited to share this?
    • Does this provide real value to my audience?
  7. Repurpose the Content: Once your series of articles is complete, use the LLM to transform them into different formats.
    • Example Prompts:
      • "Summarize this article into a 5-part X/Twitter thread."
      • "Combine the key takeaways from these last 5 articles into a webinar outline with 3 main talking points."
      • "Turn this article into a LinkedIn post of under 300 words."

Graph View

Backlinks