My son Kaiser is 9 months old. He's been on an AI-generated developmental curriculum since he was 6 months old.
I'm not a developmental psychologist. I'm an engineer who builds AI systems. When Kaiser was born, I did what every new parent does: I read the books, asked the pediatrician, downloaded the apps. It was all the same. Generic milestone checklists. "Your 6-month-old should be doing X, Y, Z." No personalization. No adaptation. No data.
That bothered me. I track everything else in my business with AI. Why would I leave my child's development to guesswork?
What I built
Prodigy started as a tool for my own family. The idea was simple: give an AI the same information a developmental specialist would have, and let it generate a daily plan.
I fed it Kaiser's profile. His age, his traits, his percentiles (99th height, 85th weight). The milestones he'd already hit and when he hit them. His mom Alyssa rates each activity and adds notes about how he responded.
The system generates 8 activities a day across 7 developmental categories. 6 come from our specialist engine, which runs a panel of 5 developmental experts (motor development, cognitive psychology, speech-language pathology, occupational therapy, behavioral psychology). 2 come from our creative engine, which takes a more unconventional, Montessori-inspired approach.
That's the daily loop. Do the activities. Rate them. Notes if anything stood out. The AI adapts the next day's plan based on the feedback.
What the data showed us
After a month, patterns emerged that we never would have noticed without the data.
Kaiser's motor development was off the charts. He was pulling to stand at 6 months. Cruising furniture by 8 months. Every motor activity got 4-5 star ratings. His body was ready for more challenge than a typical 9-month-old.
His language ratings were averaging 3.1 out of 5. Not bad. But compared to his motor scores, it was a clear gap. We weren't doing enough language-rich activities. Not because we didn't talk to him, but because we weren't being intentional about the kind of language activities that build specific skills at this age.
We adjusted. More consonant-sound games. More "serve and return" conversations where we wait for his response. More narration during activities. Within two weeks, his language ratings started climbing.
That course correction happened because of data. Without it, we would have kept over-indexing on motor activities because they were fun and the results were visible.
The curriculum changes everything
The biggest upgrade was moving from daily generation to a full 4-week curriculum. Instead of clicking "generate" every morning, Kaiser now has a structured weekly plan with themes.
Week 1 might focus on spatial awareness and object manipulation. Week 2 builds on that with more complex challenges. The difficulty ramps through the week: foundation activities Monday-Tuesday, building Wednesday-Thursday, challenge on Friday, reinforcement on the weekend.
It feels like having a developmental specialist design a program for your child. Because that's essentially what the AI is doing. It's just doing it at scale, for free, and adapting weekly.
Milestone predictions
This is the part that blows people's minds. Based on Kaiser's activity performance data, the AI predicts when he'll hit each upcoming milestone. Not just "he'll walk eventually." It says "based on his pulling-to-stand performance and cruising data, independent walking is predicted at approximately 10.5 months with 82% confidence."
Is it perfectly accurate? Too early to tell. But it gives us something to watch for, and it creates a conversation with our pediatrician that goes beyond "is he walking yet?"
What I'd tell other parents
You don't need to be an engineer to do this. That's the whole point. I built the tool, but using it takes less time than scrolling Instagram.
Open the app. Today's activities are already loaded. Pick one. Do it with your child. It takes 5-15 minutes. Rate it. Move on.
The AI does the hard part: figuring out what your specific child needs today, based on their specific history and trajectory. You do the important part: being present, playing with your child, and paying attention to how they respond.
That's it. That's the system. And the data says it works.