AI Explored Podcast Summary

Episode Title: Using Claude Projects to Develop Quality Content

Host: Michael Stelzner, Social Media Examiner
Guest: Casey Meehan (AI coach, Blazing Zebra)
Date: December 2, 2025

Episode Overview

In this value-packed episode of AI Explored, Michael Stelzner explores how marketers, creators, and business owners can supercharge their content creation workflows using Claude Projects. AI coach and YouTuber Casey Meehan shares his detailed process for using Claude’s project feature to streamline and scale content production—from initial concept to finished video and beyond. The conversation focuses on practical application, highlighting Claude’s strengths, typical misconceptions, the “few shot” training method, and a step-by-step walkthrough of Casey’s reproducible system.

Key Discussion Points & Insights

1. Casey Meehan’s Journey into AI

• Background: Ran a marketing agency focused on long-form content, fascinated with automation and AI since childhood.
• Pivots: During COVID, began leveraging machine learning for agency analytics; post-ChatGPT, pivoted to full-time AI coaching.
• Quote:

  “Within months of ChatGPT rolling out, I had my first AI coaching clients and really pivoted my agency.” (03:07, Casey Meehan)

2. Claude Projects: Addressing Misconceptions and Unique Benefits

Misconceptions About Claude

• People often see Claude as a “second-tier” LLM compared to OpenAI’s offerings.
• Contrary to popular belief, Casey (and many in Silicon Valley) identify Claude as an outstanding writing assistant, especially for marketing use cases.

Project Workspaces and Memory

• Structure: Each project is its own workspace with a chat interface, instruction box, and the ability to upload files (knowledge base).

• Advantages:

  • Retains context across chats in a project (unlike some custom GPTs that suffer “amnesia”).
  • Consultative, creative, and approachable communication style.
  • Delivers tailored, high-quality writing.

• Memorable Quotes:

  • “Claude is a fairly new tool, but it happens to be the best writer, I think, of all, all of the frontier large language models.” (06:18, Casey)
  • “With a cloud project, ... all the other chats that are inside that project are all inter-exchange into the knowledge base... you can have hundreds of chats inside of a project.” (10:24, Michael)

3. The Few Shot Training and IPO Method

The Concept

• Few Shot: Providing the AI with 3–10 concrete examples in the knowledge base, rather than elaborate prompt instructions.
• Show, don’t tell: Structure output by example rather than description.

The “IPO” (Input, Process, Output) Process

• Input: What do you give the AI? (e.g., rough ideas, transcripts, outlines)

• Process: The clear, minimal system instruction.

• Output: The desired artifact (titles, hooks, outlines, eBooks, etc.)

• Quote:

  “If you can put a few different examples of what you’re trying to create in that knowledge base, you’re going to have a very powerful tool on your hands here.” (12:56, Casey)

4. Claude Project Setup: Step-by-Step Walkthrough

4.1. Brainstorming Titles and Thumbnails

• Project contains:

  • 10 high-performing past video titles
  • Brief thumbnail descriptions (can use AI to help analyze visuals)

• System instruction: “User will provide video ideas. Your job is to generate video titles and thumbnail ideas based on the uploaded examples.”

• Chat workflow: User iteratively refines titles/thumbnails in conversation.

• Tip: Use minimal, clear instructions and focus on the examples.

  • Timestamp: 28:46 — Casey explains the set-up:

    “I have just one file that has 10 different of my best performing video titles and then a little description of the thumbnail that went with each of those.”

4.2. Writing the Video Hook

• Project contains:
  • 3 examples (documents) of hooks from top-performing videos
• Input: Finalized title and thumbnail ideas
• Output: Script for the first 30–60 seconds
• Creative diversity: Choose hooks that span style/approaches for more variety.
  • Quote:

    “You want maybe three to ten different ones... if you can have three, like pretty different ones, that would be ideal.” (36:22, Casey)

4.3. Creating the Outline

• Project contains:
  • Several past outlines, ideally diverse in style and structure
• Input: Title, thumbnail, finalized hook
• Output: Detailed video outline (manually refined as needed)
• Quote:

  “With that, I’m pretty much ready to record.” (37:06, Casey)

4.4. Video Critique Bot

• Purpose: Quality control after recording
• Project contains:
  • Transcripts of high-performing past videos
• Workflow:
  • Upload new video transcript post-recording
  • AI compares new transcript to best-practice examples; suggests improvements (structure, CTAs, pacing, etc.)
  • Quote:

    “It’ll say, ‘How does this one not match up with these others? What am I missing here?’... it will critique my performance in a way that I can go back in and edit…” (38:01, Casey)

4.5. eBook Creation

• Project contains:
  • Sample eBooks (each as a separate document) as templates for structure and style
• Workflow:
  • After final transcript, prompt Claude to draft an eBook expanding on video content, using the samples for reference
• Result:
  • Produces a 10–12 page draft (manually edited later)
  • Used as a paid product on Patreon or as a lead magnet
  • Quote:

    “With each video, I create a complimentary ebook which is over 20 pages long, which really extends everything in the video...if it wasn’t for cloud projects, I would not be able to do this.” (41:47, Casey)

5. Best Practices & Technical Tips

• System Instructions:

  • Less is more; don’t over-engineer. The more compact and clear, the better.
  • Rely more on strong, varied examples in your knowledge base.

• Knowledge Base Management:

  • Refresh with new high performers periodically (e.g., quarterly)
  • Examples should be relevant and diverse when possible

• Model Selection in Claude:

  • Experiment: Casey often prefers Opus for writing, Sonnet 4.5 for coding
  • “If it’s working for my workflow… experiment on something that’s not working, I guess, rather than something that does.” (20:53, Casey)

• Artifacts Feature:

  • Useful for editing larger outputs like eBooks
  • Turn on specifically for large creative tasks, but can become overwhelming if used everywhere (40:27–41:27)

6. Memorable Quotes

• “Claude is always thinking one step ahead ... it’s so much better at thinking through what that next step should be.” (11:34, Casey)
• “The less elaborate that [system instruction] is and the more examples you give it in the knowledge base, the better it’s going to perform.” (27:05, Casey)
• “Claude never gets tired.” (45:16, Michael)

Notable Timestamps

• 02:06 - Casey’s entry point into AI and agency background
• 06:18 - Why Claude is underrated and preferred for writing
• 10:24 - Project memory and context advantages
• 12:56 - The few shot training method and its power
• 24:16 - IPO method explained
• 28:46 - Example: Title and thumbnail generator project set-up
• 33:27 - Next steps: Creating hooks, outlines, and chaining projects
• 38:01 - Using a critique bot for post-production improvement
• 41:47 - Automated eBook generation as a value-add or paid asset
• 45:14 - Integrating Claude Projects into a weekly content engine

Resources & Guest Info

• Casey Meehan’s YouTube: @Blazing Zebra
• Website/Newsletter: BlazingZebra.AI
• Patreon: Access to eBooks, group coaching, coding cohorts, and more

Takeaways for Listeners

• Claude Projects can dramatically accelerate and quality-control your content marketing workflow—even for solo operators or small teams.
• Use clear, simple system instructions; invest your energy into curating exemplary knowledge base files.
• Build out modular Claude Projects for each step (ideation, scripting, reviewing, eBook, etc.)—allowing you to chain together scalable, reliable workflows.
• Keep examples refreshed, and experiment with Claude’s evolving models for best results.
• This “few shot” approach with Claude enables personalized, context-rich automation without sacrificing human oversight or creativity.