The Five Day Build: This is the second post in a week-long series about how to use AI in a progressive, multi-dimensional way to build more creative projects and outcomes. This week, I'm using AI to progressively build out a pilot concept for a sitcom based on a faux-real version of my life. You can read the first post here.
Welcome to Day 2 of the Five Day Build.
Yesterday I used ChatGPT to conceptualize a pilot episode plot for a faux sitcom about my life. (You can read it here.) But I had notes. I left yesterday realizing that I needed to seriously level up my game and general knowledge of understanding how sitcoms work.
So today I decided to focus on two primary next steps: Better Research and Stronger Character Card Development. Here's how things went down.
The Process:
Time Spent: 60 minutes
Artifacts Created:
Deep research piece on sitcom writing best practices
Fleshed out Character Cards
An updated structure for the pilot episode following more classic A, B, and C plot themes
Time boxing is really important on AI tasks due to the infinite and limitless potential of how to use these tools. So I decide to hold myself to a strict, 60-minute "AI Power Hour" this morning to see how far I could get on fleshing out a better version of a pilot concept.
Since I've never written a sitcom before, the first thing I had to do was teach myself a little bit about how comedies are typically structured. So I posed this prompt to Gemini Deep Research:
"i'm working on a sitcom pilot and would like to understand some of the basics about sitcom story arcs, character personas, and more. help me write a prompt to ask a deep research AI to teach me about some basic frameworks and best practices for sitcoms"
As you can see, it determined an appropriate structure for the research report on sitcom story arc best practices and researched more than 70+ websites.
While that deep research query ran, I also opened a Project in Claude to give me a place to dump all of my artifacts and build out some of the more creative elements of the story. I worked with Claude to assign some project instructions based on my goals and renamed all of the characters.
After receiving these instructions, I used the chat interface in the project itself to help me flesh out the main characters. Claude prompted me with key questions and themes related to the motivation and goals and quirks of each character, which I answered in a Google sheet.
By that point the Gemini deep research query ran its course and generated a 26-page report for me on the history and best practices of sitcoms. Then I did these things in tandem:
The Multi-Threaded AI Steams Lets You Do More At Once
I did these three things in tandem to help me move faster on next steps:
I downloaded the report and fed it to Claude as background details for the project so it could make more informed choices about characters and plot points
I asked Gemini for a summary of the report it just ran so that I could get an AI overview of the main ideas
I pulled out two main concepts from the report (the character types and the plot themes - A, B, and C) and started two simultaneous threads in Claude to help me build a more robust set of characters and a stronger story arc for our pilot episode
Once we fleshed out the characters, I downloaded that and then fed it right back to Claude as more training data. Then, to simplify how I'd explain this process to all of you, I worked with Claude to spin up quick "trading card personas" of each character, which you can see here (screenshots pasted below).
Then my 60 minutes were up, so I spent 30 minutes writing up this blog overview.
The Result: Character Cards and a Better Pilot Plot
Here's the resulting view of the final cast of characters that resulted from this mornings warpspeed brainstorm for my sitcom pilot. You can also view a more detailed view as a Claude artifact here.
The other thing I had Claude build for me was a better synopsis and plot focus for yesterday's pilot episode, given some of these new characters and a smarter conceptual approach to sitcom writing.
The Readout
Today felt like an exciting explosion of possibilities to expand this project into a more creative direction, and it was frustrating to hold myself to a strict 60-minute timer because I wanted to keep building. But I was really excited by how quickly I was able to rapidly up-level the overall plot, the characters, and get a primer on best practices for writing sitcoms.
What the AI Did Well:
Deep Research - Gemini's deep research article was gold. I now have a 26-page research paper called, "The Blueprint for Laughter: Mastering Sitcom Story Arcs, Characters, and Best Practices for Your Pilot" that I can't wait to dig into. The real unlock for me was then using the AI to extract the key takeaways so I could keep moving.
Quick Character Card Buildouts with Claude - I loved how quickly I was able to spin up a visual view of my characters to show off the cast without bogging myself down in more text. I also really appreciated Claude's prompts back to me to offer specific details that would enhance each character.
What the AI Could Have Done Better
Creative Conceptualization - The characters as they stand now still rely pretty heavily on themes, plot points, and specific details from my own life. I wish I'd been able to inject a bit more creative energy for the AI to come up with a few new ideas and character nuances on its own.
Overall Season-Long Concepts - At my prompting, we dug right into action. But I sort of wish Claude had questioned ChatGPT's original idea a bit more, rather than stick so closely to the original plot that w conceptualized. My guess is there are better pilot concepts that would make this stand out.
Lack of Imagery Details in Characters - I was disappointed to not get more visual cues incorporated into the character cards. This is an important next step, which I'll need if I want to generate any sample images or even a "teaser reel" based on this concept by the end of the week.
What's Next
Now that I've got the bones behind what makes a good sitcom, a better understanding of narrative structure, and more characters to play with, I'd like to spend tomorrow fleshing out the current draft into a more legitimate pass at what a full episode script could look like, and maybe also try to add some visual identity to the brand and characters.
What else do you think I should try next? Let's see what happens tomorrow...
Day 2 of the 5-Day AI Build This is the second post in a week-long series about how to use AI in a progressive, multi-dimensional way to build more creative projects and outcomes. On Day 1, I wrote a sitcom pilot with ChatGPT. But I had notes. On Day 2, I conducted deep research on sitcoms with Gemini and built out fully fleshed character cards with Claude See how it turned out here: https://hardmodefirst.xyz/the-5-day-build-day-2-research-and-background What should I try tomorrow???