20 Million Lives. 3 Million Nurses. One Question That Changes Everything

20 Million Lives. 3 Million Nurses. One Question That Changes Everything. Yesterday in Porto Alegre (Brazil) at the 55th annual conferene of Unimed, I was brought in by Fernanda Cunha and her stellar team (who represent my speaking activities for the Americas), together with Rafael J. Grossmann, MD, MSHS, FACS and John Nosta to paint a picture of what lies ahead. After Raphael showed how technology in 'realities' like VR, AR, and XR are evolving and what they need to take into account in terms of the speed of development, John urged them to THINK. To think about what will change with the use of AI, how de-learning and de-skilling is a serious risk. I spent 60 minutes with Unimed's 340 cooperatives asking them a question most healthcare systems won't ask until it's too late: Who owns the infrastructure that detects what's about to go wrong? Here's what I told Marcos de Almeida Cunha who moderated my session, and his 800 collaegues in the audience. "You have 115,000 cooperados. You have nearly 3 million nurses. The capacity is there. But you're leaving a second layer on the table."

Prevention that prevents. Not campaigns. Not intentions. Detection infrastructure. The kind that catches early change before it becomes crisis. Before it costs everything. Most cooperatives I work with are still building this. You have the structure to own it completely.

And then the hard part: Make sure you own the layer that sits between detection and decision. Own the inference to strengthen the prevention part. Own the data. Cooperatives are built for exactly this,collective ownership, local control, margin stays inside.

Don't let someone else build the intelligence layer and sell you the results. Don't become dependent on their data delivery, timing, or pricing models that rise with your success. The moment is now. Not because of hype. Because capital is already moving. Because the cooperados will only fund this if it's infrastructure, not if it looks like redistribution. And because once the layer is built by someone else, the cost to reclaim it is measured in years and margin. Two things have to happen in the next 90 days:

1. Define what's shared (identity, standards, validation, compute) and what's theirs (clinical decisions, community, control). No grey zone.
2. Prove it works. One clinical problem. End-to-end. Show the margin. Show the ownership.

Then vote. Then move. I've seen what happens when a cooperative really understands this. They don't just survive transitions; they lead them. Unimed Nacional has the structure. They have the scale. They have doctors and nurses where prevention actually matters. The only question left is speed. What I'm watching for next: 24-48 hour working session with their board and directors. If they move at the pace I think they will, this becomes a case study in how you actually build infrastructure instead of chasing tools.