Podcast Summary: The ChatGPT Experiment – Ep 101
"Creating Your Own Personal Review Editor"
Host: Cary Weston
Air Date: March 10, 2026
Episode Overview
In this episode, Cary Weston dives into the practical process of creating your own personalized "review editor" and scoring matrix with ChatGPT (or similar AI tools like Claude). He explains step by step how to systematize your review or editing process for repeated tasks—be it your own work or others’—and make it objective, efficient, and scalable using AI. Cary’s trademark approachable tone makes powerful AI techniques accessible to beginners and professionals alike, with real-world examples and actionable frameworks.
Key Discussion Points & Insights
Setting the Stage: Why Create a Review Editor?
- Cary describes the repetitive nature of reviewing output (like documents or proposals) and the challenge of maintaining consistency and objectivity.
- Emphasizes that most people rely on instinct rather than a deliberate checklist for reviews, but AI can help formalize and scale this process.
"Most of us don’t think about how we should do it, right? ...Most of us don’t have an official checklist—a lot of it’s just experience and instinct."
—Cary Weston (07:07)
Using AI (ChatGPT & Claude) for Review Consistency
- Cary talks about how he leverages both ChatGPT and Claude for review tasks:
- Claude for more empathetic and human-sounding writing.
- ChatGPT for brainstorming and functional analysis.
- His workflow often includes having each AI tool review the other's output for a broader perspective.
"What I find—some analytical and some empathetic kind of advantages—because it’s almost like two different brains."
—Cary Weston (05:45)
Building a Scoring Matrix: Cary’s 4-Part Framework
Cary recommends a 4-step conversational structure for setting up a review tool with AI:
- Identify What You’re Doing
- Explain Why You’re Doing It
- Define What Success Looks Like
- Invite the AI to Ask Questions
Example Application (Sales Proposal Review):
- Clearly state you want the AI to review sales proposals for consistency and excellence.
- List out key review elements or categories (e.g., clarity, persuasiveness, grammar, completeness, relevance).
- Ask the AI to score each area from 0 (poor) to 5 (excellent), aiming for 4+ in every category.
- Encourage the AI to ask clarifying questions.
Sample Conversation Outline (paraphrased):
"I create sales proposals over and over again... I want to make sure we’re being consistent and true to what needs to be done... I need to create a review editor... Here are the elements I want you to review and score from 0–5..."
—Cary Weston (09:48)
Using AI to Define Standards & Success
- Suggests giving examples of “good” work to the AI so it can interpret what excellence looks like.
- Explains the AI will often prompt you for more specifics about what makes each element “good” or “bad” and may even conduct a mini-interview to extract tacit knowledge.
"It may ask you—it did when I did this with a client—it asked me for examples of what good looks like. And that’s even better, because it can analyze what good looks like and bring that in..."
—Cary Weston (13:20)
Moving from Prompt to Project or Skill
- Once you’ve developed your criteria, you can:
- Turn it into a reusable prompt.
- Convert into a “custom GPT" (in ChatGPT) or a "Skill" (in Claude) for future use.
- Or, add as instructions to a project within either tool.
"This could be—[in Claude] they call it a skill... In ChatGPT, you might create a custom GPT... The key here is to tell the tool what you need, give it the elements, and have it work with you to pull from you what good looks like."
—Cary Weston (12:45)
Benefits: Objectivity, Efficiency, and Overcoming Bias
- The AI-based review editor helps:
- Make the process more objective ("information is no longer just in your head").
- Find aspects you hadn’t considered before.
- Reduce the "curse of knowledge"—it helps you see your work from an outsider’s perspective.
"You’re making the review process more objective and you’re making it faster, more efficient... it is helping others see objectively because we get so into the weeds, don’t we?"
—Cary Weston (15:24)
Implementation Example: How to Use Your Review Editor
- Upload your document (e.g., a proposal).
- Tell your custom AI editor to “go to work”—it reviews based on your criteria, scores each area, and suggests improvements.
"I upload the proposal, I share the proposal, and I say ‘Go to work’ ...it’ll give you the five categories you gave it, it’ll rate it, and give you suggestions on how to make it better."
—Cary Weston (14:18)
Cary’s Final Takeaways
- Focus isn’t on a prescriptive template, but on providing structure and fostering your own critical thinking using AI.
- Invites listeners to try the approach and give feedback.
Notable Quotes & Moments
-
On defining your standards:
"The goal here is to define the elements, and the range by which what bad and good look like." —Cary Weston (13:18) -
About objectivity:
"It's difficult sometimes to put yourself in the seat of the customer and look objectively at a document and act as if we don't know everything that we know. That curse of knowledge is real." —Cary Weston (16:00) -
Listener engagement:
"Let me know how you're using something like this, this already. If you are, or if you give it a go, if you have a conversation, share with me how it went." —Cary Weston (15:00)
Giveaway Announcement [17:30]
- Winner: Heather Thorsen of HT Artistry (Boston)
- Prize: Tickets to AI Business World in Anaheim, CA.
- Heather’s motivation: Wants to learn new ways to leverage AI and share with peers.
Timestamps for Key Segments
| Timestamp | Topic/Quote | |-----------|----------------------------------------------------| | 03:20 | Why we need a review editor | | 05:45 | Comparing Claude & ChatGPT as review tools | | 07:07 | Instinct vs. formal review checklist | | 09:48 | Framework for talking to AI about review criteria | | 12:45 | How to make the process reusable (Skills/Projects) | | 13:18 | Defining standards and examples | | 14:18 | Using your review editor on a real document | | 15:24 | Objectivity and efficiency gains | | 16:00 | Overcoming the curse of knowledge | | 17:30 | Giveaway winner announcement |
Tone & Style
Cary’s style is approachable, optimistic, and practical, regularly inviting listeners to share experiences and experiment without fear of making mistakes. His storytelling (local color about Maine’s “mud season” and his early boat drop-off) adds authenticity and relatability.
Listener Takeaways
- Anyone can create a personalized review tool with ChatGPT or Claude by simply structuring the conversation and defining what great output looks like.
- Such tools lead to more objective, thorough, and repeatable reviews, and help overcome personal bias.
- Start small: List your review criteria, talk it through with your AI tool, and refine as you go.
- Share your results with Cary and the community for collaborative improvement.
Final advice:
"The most important element to becoming more comfortable and efficient with ChatGPT and tools like it is your own curiosity." (18:50)
For first-time listeners:
- Cary welcomes you to explore the archives for more practical AI guidance and encourages sending in ideas for future episodes.
