AI in the NHS and Local Government: Why It Has to Be Built With Frontline Staff

‍

AI is already making its way into some of the UK's most complex public institutions but the early results are a lesson in what happens when technology is built without the people on the ground.

‍

Take the NHS, where ambient voice technology is being rolled out to automatically transcribe clinical interactions. On paper, it's a huge win: less admin time for doctors, more time with patients. In practice, AI-generated patient summaries are often too long-winded, and the key points pulled out at the top aren't always what a clinician actually needs when they're scanning a file in seconds. A summary written by a doctor follows a structure they instinctively understand.= whereas an AI-generated one doesn't because it wasn't built with that workflow in mind.

‍

Local government is facing a parallel challenge. Local government reorganisation (the redrawing of boundaries between councils) is creating major headaches around data: where it's stored, how it's structured, and how it gets transferred when the map changes. It's a problem that sits right at the intersection of technology and operational reality.

‍

‍

In short, AI can't be built in isolation and then handed to frontline staff to make it work. It has to be developed alongside the people who'll actually use it i.e. clinicians who know how a patient record needs to read, or council staff who understand how data really moves through their systems.

‍

‍