Three months ago, I did something that felt slightly voyeuristic — even though the data was entirely my own.
April 2026 update: When this was first published, the examples here came from my own texting history. Amicai now also ships on Android via getamicai.com/android, so the reflection and insight layer is no longer an Apple-only story.
I let an AI read 90 days of my text messages. Every conversation. Every "lol," every "running late," every heartfelt paragraph sent at 1 AM. Then I asked it to tell me what my relationships actually look like.
The results were uncomfortable. Not because anything was wrong, exactly. But because the data made visible what I'd been choosing not to see.
What I Expected
I expected to learn that I'm a pretty good friend. I text people back. I remember birthdays (mostly). I check in when someone's going through a tough time.
I expected the AI to confirm what I already believed about my relationships.
What Actually Happened
Finding #1: I Talk to 4 People. Everyone Else Is Fading.
Out of roughly 30 people I'd consider "important" in my life, I actively communicated with 4 of them on a regular basis. My partner, my business partner, my best friend, and my mom.
Everyone else? Sporadic at best. My college roommate — someone I lived with for three years — hadn't heard from me in 47 days. My cousin who I'm supposedly close with? 62 days of silence.
I didn't decide to let those relationships fade. It just happened. The AI showed me the gap between who I think I am and who the data says I am. These are the classic signs of drifting apart — and I had every single one.
Finding #2: My Conversations Are Shallower Than I Thought
The AI analyzed conversation depth — not just frequency. It looked at whether messages were logistical ("want to grab dinner?") or substantive ("how are you actually doing with the move?").
Turns out, 70% of my conversations were purely logistical. Even with people I love, most of our exchanges were coordination, not connection.
Finding #3: I'm Better at Responding Than Initiating
Here's a pattern I never would have noticed: I almost never reach out first. When someone texts me, I respond quickly and warmly. But I rarely start the conversation.
This means my relationships are entirely dependent on other people maintaining them. If they stop reaching out, the relationship dies — and I might not notice for months.
Finding #4: Life Events Were Slipping Through
My friend mentioned she was thinking about going back to school. My brother casually dropped that he was stressed about a health scare. My neighbor told me about a new hobby he was excited about.
I heard all of these. I registered none of them. The AI caught every one and flagged them as moments I should have followed up on — and didn't.
What I Changed
I'm not going to pretend this was a dramatic life transformation. But three specific things shifted:
I started initiating. One text per day to someone I haven't talked to recently. It takes 15 seconds. The responses have been overwhelmingly positive — people are surprised and happy to hear from me.
I started noting details. When someone mentions something important, I make a point to remember it and bring it up later. "How did that doctor's appointment go?" is a small sentence with enormous impact.
I accepted the gap. I can't maintain 30 deep relationships simultaneously. But I can be intentional about which ones I invest in, rather than letting chance and proximity decide for me.
The Uncomfortable Truth
We all have a story about who we are in our relationships. The data tells a different story. Not a worse one — just a more honest one.
The question isn't whether you're a good friend or a bad one. It's whether you're paying attention to the relationships that matter, or just assuming they'll maintain themselves. (And if you're wondering what happens to your data during an experiment like this, here's what AI companies actually do with it and how your phone numbers stay protected.)
They won't.