How Domain Experts are using AI
Real stories from the field of people turning expertise into AI-powered advantage
While everyone debates whether AI will replace jobs, there’s a quiet revolution happening. Domain experts, from farm-based newsletter analysts to solopreneurs with no technical background, are not waiting to be disrupted. They’re using AI to turn their knowledge into products faster than ever before.
And it’s working better than anyone expected.
The New Building Reality
Two years ago, if you wanted to build software, you had two options: learn to code or hire someone who could. Both took months and thousands of dollars.
Today? I just watched a filmmaker ship a video platform in three months while traveling Thailand. A copywriter built semantic search tools from her Arkansas farm. A retiree finally created the niche directory they’d dreamed about for decades.
The shift is simple: We’ve moved from “learning to code” to “learning to collaborate with AI.”
Here’s what this looks like in practice. When I built Quick Viral Notes, I didn’t write a single function from scratch. Instead, I used Cursor (AI-powered IDE) to help me build what I needed. My process:
Day 1: Sketched the concept on paper, fed it to Claude
Day 2-3: Generated Python scripts to run through the complete logic
Day 4-5: Setup the complete frontend and backend to handle the core functionality
Day 6: Hosted for free on Oracle’s lifetime free plan
Day 7: Shared publicly
Total cost: $20 Cursor subscription + trickles of Claude chatting
Total time: 7 evenings
Result: A working product that turns long-form articles into 18 note ideas, now with hundreds of paying users
Look, I’m not going to pretend this was flawless. The second day it launched, it got stung because of the traffic. From time to time I received bug reports, and the UI looked really ugly on mobile. But: it worked well enough to solve MY problem. And that’s where every good product starts.
The emergence of what I call “Technical Product Builders” — people who combine domain expertise with AI collaboration skills — is creating opportunities that didn’t exist a year ago. The competitive advantage isn’t technical mastery anymore. It’s understanding problems worth solving and having the courage to build solutions.
Builders Who Prove Anyone Can Do This
I keep seeing this pattern emerging, in my own building and in other builders I follow. But most of these builders were working in isolation, sometimes discouraged by “mainstream” developers, and don’t have means to learn from each other.
So I built vibe coding builders — a simple directory where these builders can list themselves and showcase their product. Every Friday, I feature one of these builders in my newsletter, understand how they’re actually doing it, and help them get discovered.
From those conversations, here’s exactly how different people are making this work, with real numbers and workflows.
Industry Disruptors Who Saw Change Coming
Kenny: The filmmaker who adapted
Kenny spent 13 years as a director of photography, shooting for big tech conferences. Then AI video tools started getting good, really good. Instead of panicking, he got curious.
“On July 5th, my wife and I were having coffee in Chiang Mai, and I was dreading having to renew my annual Vimeo subscription at $130. I literally said, ‘Hmm, why don’t I try building an alternative?’”
His tactical approach:
Week 1: Created mockups in Canva (free)
Week 2-4: Used Replit’s Assistant mode (5¢/request vs Agent mode at $$$)
Month 2-3: Iterative building while traveling
Key workflow insight: “I create a dedicated folder in ChatGPT to serve as my evolving working document. The chat retains all your questions within that folder, even across new chat sessions!”
Results: 20 active users, $25/year pricing model, freemium approach Current status: Scaling ProudWork as a creator platform
What struck me about Kenny: he didn’t wait until he “knew enough.” He started building because he had a problem that annoyed him daily. Sometimes that’s all you need.
Technical People Who Embraced Collaboration
Finn: The data expert who 10x’d his output
Finn analyzed 1.3 million Substack Notes while juggling a demanding day job. But despite being deeply technical, he readily embraced AI assistance.
His setup:
Primary tool: Cursor for building, ChatGPT for brainstorming
Stack: Chrome extensions (zero hosting costs)
Workflow: Build first version with AI, then iterate manually
The numbers that matter:
Revenue: $10,000+ from Substack tools
Time investment: ~4 hours per week (evenings/weekends)
Development cost: Essentially free (extensions don’t need servers)
“I have some Gumroad traffic from perplexity.ai but the interesting detail was that the conversion rate was 50%. Normally my conversion rate is between 4% - 17.2% from other sources.”
Finn’s key insight: AI amplifies expertise, it doesn’t replace it. His domain knowledge of creator analytics + AI development speed = products people actually pay for.
Knowledge Workers Who Discovered Building
Karen: The “why not try?” copywriter
From her 143-acre farm in Arkansas, Karen tackled vector databases and semantic search. No computer science background. Just curiosity and users who needed better newsletter discovery.
“Well, perhaps I just don’t know how difficult some things are, so I’m willing to try them!”
Her evolution:
Started with: Copy-paste coding for a year (slow but educational)
Switched to: Claude Code (“I’m 10 times faster with it”)
Current setup: Claude Code for rapid development, careful version control
The breakthrough moment: A user suggested semantic search. Instead of dismissing it as “too technical,” Karen spent a weekend learning vector databases.
“I’m using sentence transformers to give each article a similarity ranking from 0 to 100. Most results on a digest show between 50-100% similarity based on the semantic search.”
Result: StackDigest with enterprise-level search capabilities built by someone who learned databases from YouTube.
That “why not try” mentality? That’s the real unlock. Not technical skill, just willingness to attempt things that sound impossible.
The Pattern Behind These Stories
None started as “builders.” They started as people with problems only they understood from their daily work. Kenny saw disruption coming and pivoted instead of resisting. Finn amplified his expertise rather than competing with AI. Karen stopped saying “that’s too technical for me” and just tried.
What gets me most excited: I’ve talked with retirees in their 60s and 70s who are building the specialized directories they’ve dreamed about for decades. Professional expertise meeting accessible AI collaboration. One told me: “I’ve wanted to build this for 20 years. I finally can, and the process itself is the most fun I’ve had in years.”
The shift isn’t about becoming technical, it’s about recognizing that daily professional problems are actually building opportunities waiting to happen.
The Real Numbers
Let’s be honest about what this actually costs and takes:
Time investment across our builders:
Kenny: 3 months part-time while traveling
Finn: 4 hours/week evenings + weekends
Karen: Evenings after full-time work
Tool costs (monthly):
Cursor: $20/month
Claude Code: $0-50+ depending on usage
Replit: $0-20/month
Hosting: $0-100/month depending on your needs
Success timeline reality:
First working version: 2-6 weeks
Something people use: 2-4 months
Revenue (if pursuing): 6 months - 2 years
Sustainable income: Highly variable
Failure rate: High. I’ve built plenty of things no one wanted, and most builders go through multiple failed attempts.
The difference: Each failure teaches you something specific about what people actually need.
Now that you know what to expect, here’s exactly how to start:
Your Next Play
Start Here: Find Your Building Opportunity
Don’t just “list problems.” Use this exact framework:
Week 1 assignment: Every time you think “I wish there was a tool for this,” write it down. Set a phone reminder for 6pm daily: “What annoyed me today that software should solve?”
End of week: Which frustration came up most? That’s your project.
Example from my own log:
“I wish I could quickly analyze my Substack’s performance vs others” (5 times)
“These newsletter tools don’t understand my specific needs” (8 times)
“I can’t easily track which builders I’ve discovered” (12 times)
Winner: Builder discovery problem → vibecoding.builders
Validation filter: Would you pay $20/month to solve this problem for yourself? If not, find a different problem.
Build It: Create Your First Solution
Month 1 plan:
Week 1-2: Choose your tool
If you can afford $20/month: Cursor (best for iterative building)
If you prefer pay-per-use: Claude Code ($2-25 per project)
If you want free starting point: Replit (free tier, paid features as needed)
Week 3-4: Build the simplest possible version
Don’t add features
Don’t make it pretty
Make it work for YOU first
My first-build checklist:
Solves my specific problem
Works on my computer
I would use it tomorrow
Takes <30 minutes to test
Launch It: Get Your First Users
Week 5-6: Share what you built
Post about what you’re building. Not the code, the problem you’re solving. “Building a tool to help [specific people] with [specific frustration] because [personal reason].”
Launch tactics:
Share in relevant communities where people have this problem
Ask 5 people with the same frustration to try it
Post on your social media with the problem statement
Don’t focus on features, focus on the pain point you’re solving
Success metrics for launch week:
3+ people actually try your tool
1+ person says “this is useful”
You identify the biggest remaining friction points
Grow It: Connect and Scale
Month 2+ tactics:
Community strategy (track this stuff):
Join one Discord/Slack community focused on builders
Comment meaningfully on 3 builder posts per week
Share one thing you learned each week
Help one person solve a problem you’ve already solved
Feedback collection:
Give your tool to 5 people who have the same problem
Don’t ask “Do you like it?”
Ask “What’s still annoying about this?”
Build what they tell you, not what you think they need
Revenue validation (if that’s your goal):
Would 10 people pay $10/month?
Or 100 people pay $1/month?
Or 1 person pay $100/month?
Build for the model that matches your energy level
Expansion signals:
People share your tool without you asking
You get feature requests from strangers
Someone else tries to copy what you built
People ask “can you build this other thing?”
What’s Really Happening
We’re witnessing a fundamental shift in who gets to create solutions. The barriers that kept professionals from building software are collapsing.
The old model: Have idea → Find technical co-founder → Raise money → Build → Hope
The new model: Have problem → Build solution → Share → Iterate → Grow organically
The competitive advantage isn’t technical expertise anymore—it’s understanding problems worth solving and having the courage to build solutions.
So question for you: What daily frustration are you going to stop accepting and start building a solution for?
 | A guest post by
|
Thank you Jenny for deep diving into these very tangible examples, it definitely expanded the possibilities for me :)