10 Things I Learned in 2025 Building a Healthcare AI Product
Lessons from a year of doubt, growth, and unexpected wins.
I spent 2025 improving and growing SOAPNoteAI - a healthcare documentation tool used by clinicians across multiple specialties. It was the hardest year yet. Here’s what it taught me.
1. Balance might be a myth. You can’t half-commit.
I left Amazon thinking I’d own my time and balance work and family. 2025 taught me this: when you’re competing against well-funded teams shipping weekly, you can’t half-commit.
The freedom I wanted required more from me, not less.
2. Doubt doesn’t come once. It comes in waves.
“Should I just go back to a job?”
I asked myself this multiple times in 2025. After churn spikes. After competitor announcements. Sometimes for no reason at all.
It’s not one crisis. It’s a recurring question you learn to live with.
3. What gets you through: honest conversations and customer wins.
Talking with my wife when the doubt gets loud. A message from a customer saying the product gave them their evenings back.
These moments don’t eliminate doubt. They give you enough fuel to keep going.
4. Curiosity can become a trap.
I wanted to learn everything. Chrome extensions, iOS, SEO.
But I didn’t need to know how to build everything. I needed to know what was best for my customers.
Curiosity without focus is just procrastination in disguise.
5. I couldn’t do everything. That wasn’t weakness. It was math.
Growth stalled. Roadmap frozen.
Ironically, I delegated naturally at Amazon. But with my own product, I couldn’t let go.
The breakthrough wasn’t motivation. It was exhaustion. I finally hired help.
6. The easiest features to add become the hardest to maintain.
Some features take two days to build and two years to support. Every convenience for users adds complexity for you.
I learned this the hard way. More than once.
7. Every customer conversation teaches you something.
A customer was close to churning. Her notes weren’t formatting correctly. Frustration in her emails.
I reached out personally. Fixed it. Explained why it happened.
She stayed. I learned exactly where my product was broken.
8. One message can inspire an entire product.
A blind massage therapist messaged me. She explained how audio dictation changed her documentation workflow.
That single message inspired our entire iOS app.
9. Well-funded competitors ship weekly. Pick your battles.
You can’t match every feature. Every price. Every marketing campaign.
Focus on what you do best. Let the rest go.
10. The unexpected win wasn’t a feature. It was collaboration.
I started 2025 wanting to do everything alone.
I ended it working with a small team—discussing problems, brainstorming solutions, moving the roadmap forward together.
I didn’t plan to build a team. It became the highlight of my year.
Final Reflection
2025 was hard. Growth slowed. Doubt was loud.
But I learned more this year than any year at Amazon.
What did 2025 teach you? I’d love to hear—reply to this post or drop a comment below.
Thanks for reading. If you’re building something too, I hope this was useful. Here’s to 2026.
Happy New Year!
— Kunal