Implementing AI in hospitals: the tech isn’t the hard part

2 minute read


Bringing the organisation along with it is the real transformation.


Most conversations about AI in healthcare focus on the technology – the model, the algorithm, the performance metrics.

But when you start implementing these systems inside hospitals, something becomes clear very quickly: AI transformation has two challenges.

The first is the data platform – making sense of fragmented electronic health record data, connecting systems that were never designed to work together, and building the infrastructure that allows analytics and AI to exist at all.

But in healthcare, those technical challenges sit inside an already complex environment of devices, workflows, and clinical decision-making.

It’s only half the story. The second challenge is changing the system around the platform.

Healthcare is a human system built on workflows that have evolved over years. Clinicians develop workarounds, each team builds their own informal unique processes, and operational rhythms emerge around how work actually gets done on the ground.


So before building technology, much of the work is simply understanding the system – where clinicians actually lose time, which tasks are manual but invisible, and which workflows exist only because the system forces them to exist.

Often the most productive sessions for us aren’t about presenting solutions. They’re about observing, listening, and letting teams redesign parts of their own workflow.

Often that conversation is enough for teams to pause and reconsider processes that have become routine.

Because building the platform is a technical challenge, but bringing the organisation along with it is the real transformation.

For those implementing AI in hospitals, what has been harder in practice – the technology or the workflow change?

Dr Caitriona Ryan is a medical doctor working across AI-enabled clinical documentation, healthcare data analytics, and enterprise AI platform implementations, supporting the development and validation of healthcare AI systems in regulated environments, in the United Arab Emirates.

This article was first published on Dr Ryan’s LinkedIn feed. Read the original here.

End of content

No more pages to load

Log In Register ×