The Future With Agentic AI in the Hospital
On Monday I explained how we are headed towards a world, where clinical work increasingly happens within a single agentic interface, which connects to and interacts with all the other systems and apps. On Tuesday I concluded that despite this unprecedented aggregation of both clinical data and intelligence, AI agents will remain at the advisory level, leaving the final treatment decision to the doctor.
There are a couple of takeaways to consider here:
- Clinicians are not going anywhere and they will be spending increasingly more quality time with the patients and increasingly less time switching between contexts and interfaces.
- Related to that, many software companies will lose their brand recognition and identity, as they will become faceless IT infrastructure, where their UI and UX play no role and their technical capabilities are to a certain extent commoditised.
- The clinical workflow will change, partly driven by automation and partly driven by changes in how clinicians work together. Right now work is done in a single-player mode. We are going to see the rise of more collaborative, multi-player mode approaches, similar to how it is in tech.
- Likewise, with all the noise removed, the patient journey could take the stage and so could a personalised AI patient manager (similar to a project manager) who’s job would be to oversee the process - from diagnosis, to treatment, to monitoring - making sure things go well.
But we shouldn’t forget that all of those agents are just software that we will need devices for - to access the agents and for the agents to capture the data. So while it seems like pure software companies will take a hit, hardware companies could develop durable moats.
The rules of the game are changing, and while we can debate about the specifics, but the direction is clear.