The system-level unlock
Everything we covered in previous sections happens within the boundaries of individual organizations. Leaders can control their own workflows, their own tools, their own change management. What they can’t control is the system those organizations sit inside.
That system has a structural problem. And it’s the one that may matter most.
"With AI and the ability to potentially anonymize data and make it more available, in two years we should be able to create systems that are more accessible and digitally forward on behalf of consumers."
Data that could improve patient outcomes sits locked inside proprietary systems, unavailable to the patients it belongs to or the providers who need it. This isn’t a new problem. What’s new is the possibility of a technical path around it.
At Optum Rx, that path runs through agent-to-agent interoperability – AI systems that can communicate directly with each other across organizational boundaries.
“When you, when you have a health plan app, you want it to have an understanding of all your benefit, all of your plan, whether it's medical, pharmacy, dental, et cetera. If we don’t have A2A (agent-to-agent protocol), you wouldn’t get that. You would just get the medical view. I want to give the health plan app pharmacy superpowers. We can do that with A2A.”
The enrollment process is another place where the system-level problem is acutely visible, and where the fix requires coordination beyond any single organization’s walls.
"We should soon be at a point where there could be some sort of standard data exchange for the fields that are always included on an enrollment form, just arranged differently – and to simply be able to have the provider push permission to transmit those data fields to the manufacturer's hub. Not even be talking about an actual enrollment form, and certainly not one that has to be faxed."
The vision underneath both of these ideas is the same: a patient who moves through the healthcare system without having to reintroduce themselves at every door. Without having to re-explain their history, re-submit their information, or re-navigate a benefits structure they’ve already navigated before.
George Van Antwerp of Prime Therapeutics envisions it as a virtual advocate.
"Almost like a digital twin of you that understands your choices, understands your plan design, understands how you make trade-off decisions – but also can tap into clinical knowledge, can tap into other sources of information. So at any time of day, any place, I can ask a question and get a real-time answer that reflects my clinical reality, my plan design, is optimized for where I am in my life, where I am in my budgeting, and can also begin to nudge me. To me, when we get there, it's going to be so powerful."
That vision is still aspirational. But it’s directionally consistent with where the technology is heading, and with what patients are already demanding, with or without the industry’s blessing.
At Infinitus, we think about the system-level unlock in terms of what becomes possible when AI connects the dots that either aren’t getting connected or require a person to connect.
The question that remains is what all of this adds up to, and how soon. In the next section, we’ll cover where healthcare AI is headed, and what the road there actually looks like.
