Wow...I'm blown away by the insights I got using these prompts and the process you laid out.
I've been building the StackContacts CRM tool that collects all my Substack data, including all DMs, emails, comments, notes, subscriber events and details, my Posts & Notes, and all my Gumroad data (products, sales, etc). All this data is stored in a local SQL database that AI can access using the MCP server. This is about 214 Mbytes of highly structured and queryable data.
STEP 1. Produced High-Velocity content and Community Pain points that I'm (painfully) aware of from these hundreds of DM threads, comments, and other data. No surprises there, but it was great to see a concrete application for the CRM tool I've been building over the last two and a half months.
STEP 2: Created a clear "Diagnostic Brief" summary and thematic clusters that all felt familiar.
STEP 3: The five layers - SCAR, PATTERN, HERESY, SIGNAL, FILTER - were the most mind-blowing part of this process. I spent about 30 minutes on this, trying to distill my experience into concise sentences.
STEP4: Wow...just wow. The content ideas are truly exceptional. The Core arguments refer to actual signals from real people in my CRM tool and address the gaps I know, but never figured out how to write about them.
STEP 5: This is the final test - Can I defend these claims in the public forum?
Congratulations on this post!
I wish I could package this process into an easy-to-use tool that could leverage the actual data, like I'm working with StackContacts CRM. This is amazing stuff.
The prompts in this post worked well with Claude Sonnet 4.5, which I used in conjunction with the Cursor app, connected to the DuckDB SQL database that contains all the details, utilizing the MCP server.
Here are some more details of the AI project that I'm currently working on:
Reading about the Backwards Law is so useful, especially we have to shape our own 'Perspective Stack' as you mentioned. Thank you Mia and Sam for this piece! 🤗Very useful and some great frameworks that incorporate the intentional pauses that every content needs in ideation and creation.
Also, you have to tell me how you generate images for your articles, Mia!😅 They seem to exhibit the perfect balance between humour and creativity!
Never giving away my visuals secrets haha 🤣 just kidding, I have a very simple 3 steps process:
1. I have an "aesthetic" prompt - using my brand colors, some combo of nostalgia vibes, film noir and artistic flairs. I don't describe anything at this stage.
2. I use another prompt to identify the best visuals placements in a newsletter but I usually generate 5-6 photos for workflows and a featured photo
3. I then use the "aesthetic" prompt and describe a scene and the core concept of it and i think of a few creative angles of how I can combine that narrative scene with the actual concept / framework.
And I use Imagen 4 Ultra to generate 4 versions for each image :)
Yes, in music, the silence is just as important as the noise... otherwise it would be just one continuous note and not so beautiful. The silence makes the magic happen.
Truly enlightening! This intentional pause framework makes AI feel like a thoughtful collaborator, not just a fast tool. Deliberate pauses in AI workflows transform speed into clarity, letting your unique perspective surface instead of just echoing what’s already out there.
This is truly some amazing work, Mia and Sam. I am without words. You really created something ingenious, non-obvious and actually useful. Please stop the internet for a week. I need to digest this to the end.
That is a very cool process, and it's no small task to create an article that explains it in a way the rest of us can understand. Awesome! I'm going to carve out some time to try this and see where it goes.
Thank you Kuet, as always! :) it's so good to hear that this was explained well, I usually struggle at making things simple haha. Lucky me, I had Sam by my side for it! 🙏
Couldn't agree more. This peice is so insightful. The idea of intentional pauses in AI is crucial for human-centered tech. Thank you for highlighting this important perspective!
I was planning my day and just set aside 90 minutes today to laser focus on this newsletter and play with the results. Solid from start to finish, well done.
Wow...I'm blown away by the insights I got using these prompts and the process you laid out.
I've been building the StackContacts CRM tool that collects all my Substack data, including all DMs, emails, comments, notes, subscriber events and details, my Posts & Notes, and all my Gumroad data (products, sales, etc). All this data is stored in a local SQL database that AI can access using the MCP server. This is about 214 Mbytes of highly structured and queryable data.
STEP 1. Produced High-Velocity content and Community Pain points that I'm (painfully) aware of from these hundreds of DM threads, comments, and other data. No surprises there, but it was great to see a concrete application for the CRM tool I've been building over the last two and a half months.
STEP 2: Created a clear "Diagnostic Brief" summary and thematic clusters that all felt familiar.
STEP 3: The five layers - SCAR, PATTERN, HERESY, SIGNAL, FILTER - were the most mind-blowing part of this process. I spent about 30 minutes on this, trying to distill my experience into concise sentences.
STEP4: Wow...just wow. The content ideas are truly exceptional. The Core arguments refer to actual signals from real people in my CRM tool and address the gaps I know, but never figured out how to write about them.
STEP 5: This is the final test - Can I defend these claims in the public forum?
Congratulations on this post!
I wish I could package this process into an easy-to-use tool that could leverage the actual data, like I'm working with StackContacts CRM. This is amazing stuff.
Wow I loveeeeeeee to hear this!! Honestly.
And what a huge project! So happy this framework helped ❤️ (and hopefully, you took enough pauses while working through it haha)
The prompts in this post worked well with Claude Sonnet 4.5, which I used in conjunction with the Cursor app, connected to the DuckDB SQL database that contains all the details, utilizing the MCP server.
Here are some more details of the AI project that I'm currently working on:
https://finntropy.substack.com/p/why-im-building-a-crm-that-remembers
Checking this out!
Reading about the Backwards Law is so useful, especially we have to shape our own 'Perspective Stack' as you mentioned. Thank you Mia and Sam for this piece! 🤗Very useful and some great frameworks that incorporate the intentional pauses that every content needs in ideation and creation.
Also, you have to tell me how you generate images for your articles, Mia!😅 They seem to exhibit the perfect balance between humour and creativity!
Yaaay thank you :)
Never giving away my visuals secrets haha 🤣 just kidding, I have a very simple 3 steps process:
1. I have an "aesthetic" prompt - using my brand colors, some combo of nostalgia vibes, film noir and artistic flairs. I don't describe anything at this stage.
2. I use another prompt to identify the best visuals placements in a newsletter but I usually generate 5-6 photos for workflows and a featured photo
3. I then use the "aesthetic" prompt and describe a scene and the core concept of it and i think of a few creative angles of how I can combine that narrative scene with the actual concept / framework.
And I use Imagen 4 Ultra to generate 4 versions for each image :)
Yes, in music, the silence is just as important as the noise... otherwise it would be just one continuous note and not so beautiful. The silence makes the magic happen.
Bella Bartok would disagree haha (composer - if you're curious you can check some of his pieces). He barely "pauses"! 😆
Truly enlightening! This intentional pause framework makes AI feel like a thoughtful collaborator, not just a fast tool. Deliberate pauses in AI workflows transform speed into clarity, letting your unique perspective surface instead of just echoing what’s already out there.
Thanks Suhrab, would love to see what this looks like in practice for you as well. 🙏
Thank you for reading Suhrab! :)
Wow.
This is truly some amazing work, Mia and Sam. I am without words. You really created something ingenious, non-obvious and actually useful. Please stop the internet for a week. I need to digest this to the end.
Yaay thank you Robin! Loved having Sam by my side for this piece!
BRB chatting to the "internet" to take a pause for you haha 😂❤️
Fantastic tutorial! Blown Away !!
Thank you Chris, appreciate it so much! 💖
Thanks Chris! 🙏
Brilliant work again. Mia and Sam, you consistently bring the kind of thoughtful depth that makes Substack worth reading.
Wow thank you Karo! Means a lot ❤️❤️
Thanks so much Karo! 🙏
Definitely the most engaging and effective piece about working with AI I've read so far. Maybe the one and only how-to useful 😉
Think I'm gonna stop every reads for one month and just focus on this.
Thank you Antonia!! ❤️
Wow. Thanks Antonia. 🙏
That is a very cool process, and it's no small task to create an article that explains it in a way the rest of us can understand. Awesome! I'm going to carve out some time to try this and see where it goes.
Thank you Kuet, as always! :) it's so good to hear that this was explained well, I usually struggle at making things simple haha. Lucky me, I had Sam by my side for it! 🙏
I struggle with the simplification myself, but this is great.
Couldn't agree more. This peice is so insightful. The idea of intentional pauses in AI is crucial for human-centered tech. Thank you for highlighting this important perspective!
Thanks so much Roxy. 🙏
Thank you Roxy! :)
I was planning my day and just set aside 90 minutes today to laser focus on this newsletter and play with the results. Solid from start to finish, well done.
Would LOVE to hear your insights after! Thank you so much 🔥
Likewise! Thanks so much Richard. That is a really decent amount of time to dedicate to this and we really appreciate it. 🙏
It's valuable. I can see it. Now I want to feel it…
Guaranteed, Mia. I'm looking forward to it ☺️
Love it. This is both profound and genius. I couldn’t expect less from the two authors.
Thank you Juan, what a lovely thing to say! :)
Thank you so much Juan! 🙏
Wild how the best ideas show up only when you stop demanding them.
Loving this ‘designed pause’ mindset 🙏
My AI is obsessed with sitting in the silence, and the meaning that is embraced in that moment.
That's a really cool AI! 😉
Absolutely loved this breakdown.
I’ve seen how pausing transforms raw ideas into something that actually sticks.
In my work with My Accountability Partner, slowing down consistently is what turns scattered notes into actionable content.
How do you usually capture those quiet insights before they slip away?