Finding a daycare in NYC is brutal. That's not an exaggeration — it's one of the most stressful things I've done as a parent.
So here's what I did. I went on Google. I checked Yelp. I looked at Google Maps. You know, the normal stuff. And every single option I found was expensive. We're talking $3,000 a month, minimum. Before application fees. Before registration fees. Before whatever else they decided to tack on.
And they were all full. Waiting lists of three to six months.
Here's the thing I noticed: the daycares that advertise — the ones showing up at the top of Google, the ones with the pretty websites, the ones with five-star Yelp reviews — they're expensive because they can afford to be. They have marketing budgets. They invest in SEO. That costs money, and who pays? We do.
I thought there has to be another way.
So I built an agent
I work on AI agents. It's what I do for a living. And I figured — why not use what I build for my own problem?
I created an instruction file. Basically a prompt that said:
- Here's where I live (specific neighborhood in NYC)
- Here's how far I'm willing to walk
- Here's my kid's age and any details that matter
- Here's what I'm looking for (full-time, part-time, whatever)
- Here's my budget
Then I let it run. For about 30 minutes, the agent did its thing:
- It used web fetch to pull public records — NYC has a lot of these, you just have to know where to look
- It used browser automation to search government databases, cross-reference addresses, check licensing records
- It refined its search based on what it found, narrowing down to places that matched my criteria
This wasn't magic. It was just code. A lot of code, actually — but the gist is: it searched everywhere, all at once, in ways that would take a human weeks to do manually.
What it found
A daycare that wasn't advertised anywhere.
Not on Google. Not on Yelp. Not on any parenting forum. The only reason it existed in any digital form was because it was listed in New York City public records.
Most of their business came from referrals. They didn't need to market. They were always full — by word of mouth alone. But here's the crazy part: they had availability. They just never bothered to update their Google listing or create a website.
I called them.
We went to visit the next day. It felt different from the other places — warmer, more personal. The woman running it sat down with us, talked to us about our son, showed us around. No sales pitch. No application fee just to be considered.
And the price? Way below what I was budgeting. I'm not going to say exactly how much, but let's just say I went from expecting $3,000 to paying significantly less.
Was I lucky? Absolutely.
This could have easily not worked. The agent could have come up empty. The daycare could have been terrible. I could have called and never heard back.
But here's what I'm not surprised about: this daycare wasn't advertised. They didn't spend money on marketing. They didn't need to — they were full through referrals. So when they had a spot open up, it didn't go onto any website. It went to whoever happened to know someone.
The fact that they were cheaper makes sense too. They weren't trying to recoup marketing costs. They weren't optimizing for search rankings. They were just... running a daycare.
I'm not saying every parent needs to build an AI agent to find childcare. But I am saying: the "normal" way of searching — Google, Yelp, whatever — shows you the options that can afford to be found. The hidden ones, the referral-only ones, the ones that are actually affordable? You won't find them that way.
Sometimes you just need to search differently.