Every family who walks into a neurology or psychiatry clinic is asked to accept something that would be unacceptable almost anywhere else in medicine: that finding the right treatment will take several attempts, that each will take weeks or months, and that no one can say in advance which will work.

My co-founder, Dr. Dilip Karnik, has spent more than forty years in pediatric neurology in Austin. He will tell you the hardest conversation in the clinic is not a difficult diagnosis. It is telling a parent that the first medication did not work, and that the next one is also, in the end, a considered guess.

I have spent three decades building data and AI systems, and four companies. When I turned to this problem I expected to find the science was not ready. That is not what I found.

The science has moved. We can read a patient’s genome, model their neurobiology, and rank therapies with real precision. The other half is missing. In the largest trial of genomic prescribing in depression, giving physicians a genetic report changed what they prescribed dramatically — and barely changed whether patients got better. The information helped at the moment of the decision. Nothing came back afterwards to say what happened.

That is the gap POD Health was founded to close.