A question I’ve asked myself everyday this week is: How do I make my projects AI-proof?
What type of information is in-demand, but AI still won’t be able to answer?
And what type of websites will still be around for years to come?
I actually think there’s a few answers here (will be making a video on this soon), but one thing I can’t stop thinking about is the power of specialized data.
Specialized data is just information that’s hard to get.
It’s information that you own, that nobody else has, or that exists in such low-quality form elsewhere that it’s pretty much useless.
When I built my thrifting directory two years ago (which still earns around $1,500 to $2,000 a month passively), I didn’t realize that I created specialized data with my $19 digital product.
And the only way I got this data was through doing unscalable things.
At the time, I called 150 stores. I spoke directly with managers. I asked questions to extract store-specific information. Then packaged it into a PDF.
But from a future-proofing angle, what I had done was create:
- Highly in-demand information
- A product no one (in their right mind) could replicate
- Information that LLMs and AI search engines could never provide
Back then, I felt kind of silly. I spent weeks collecting data for a low-priced product.
But two years later, that data is still one-of-a-kind. Most importantly, it’s invisible to Google and LLMs because it lives in PDF form behind a pay-wall.
Unscalable Tasks are the New Moat
I believe there are thousands of opportunities to acquire specialized data.
But most specialized data comes from unscalable tasks.
Everyone these days want more automation. More AI agents to automate their workflows and replace manual tasks.
Ironically, I think we’re living in a time where unscalable tasks to acquire quality data is more valuable than ever.
When you do something unscalable, you unlock data that no one else is willing to work for.
That’s what gives you leverage in an age where everything is being scraped and synthesized by AI.
You don’t have to cold-call 150 stores like I did, but you do need to find the gap between what people want and what’s hard to find.
That’s where the opportunities live.
If you don’t want to do unscalable tasks, the other way to get specialized data is to put yourself in a position where you naturally have access to it.
But it starts with building the base of your directory first.
I did this with my plasma donation directory, where I crowdsourced my specialized data around “plasma donation pay” once I had thousands of monthly visitors.
Again, no one owns this data except me!
But I put myself in the unique position to get it in the first place.
I had to earn attention before I could collect insight.
And that insight is what led to a product people actually wanted.
Opportunities show up once you’re knee deep in your niche
If you’re thinking about starting a directory but wondering whether AI is going to make it pointless, my advice is to just keep building.
You won’t discover specialized data opportunities unless you’re already immersed in the niche.
Once you publish your directory and you seek feedback (i.e. by posting on reddit), that’s the only way to start noticing what information people are hungry for and where the gaps are.
Becoming AI-proof doesn’t mean ditching SEO or abandoning what works.
It just means layering in a new level of strategy that only gets unlocked once you’re getting traffic + deep in the weeds of your niche.
So if you’re just starting out, don’t overthink it.
Build the simple version first. Let people use it. Listen to what they wish it had. Then go get that information, even if it’s messy or manual or slow.
Because that’s where I believe the moat is.
There are still thousands of underserved pockets of the internet. And most of them are invisible to AI.
SYD Pro Pricing Update
Just a quick heads up, I’ll be raising the price of Ship Your Directory Pro, my private directory and SEO community, on June 1st, 2025.
We have 135 members right now. I want to keep things as valuable as possible, which means going deep into personal projects.
Feel free to reply to this email if you have any questions. 🙂
Keep Building,
Frey
