Dr Harry Thirkettle, director of health and innovation, Aire Logic
If you have been to any health tech conference in the past couple of years, you already know the version of the NHS AI conversation that dominates the room.
It is about diagnostics and imaging, about the algorithm that catches the thing a tired clinician missed at four in the afternoon, and it is genuinely exciting work that deserves the attention it gets.
It is also, on the evidence of this survey, not what the people using and delivering NHS digital services say they need most.
We asked 50 senior NHS digital and clinical leaders a fairly blunt question: over the next five years, where do you expect AI to make a real difference in your organisation?
Not where the funding announcements are pointing, and not where the headlines are, but where they honestly think the change will show up.
Nearly half of them said the back office, which is three times the number who pointed to clinical decision support and diagnostics.
I think this is an important finding that merits close attention, because the NHS spends an extraordinary amount of its time and money on work that never touches a patient directly, from rostering and procurement to referral administration, estates, finance and the endless reconciliation of one system’s version of the truth with another’s.
If national strategy is pointed at the operating theatre while the leaders on the ground are quietly pointing somewhere else entirely, that is a misalignment we need to fix.
The second theme is one people tend to talk around rather than about.
78 per cent of the leaders we surveyed are worried about the NHS’s growing reliance on US-based technology providers for critical infrastructure, and half named data sovereignty and patient data security as their single biggest concern about international big tech.
What strikes me about that is not the strength of feeling so much as who is expressing it. This is no longer a fringe position held by a handful of purists at the edge of the debate; it has become the mainstream view among the people who are accountable for keeping NHS systems running.
Alongside that, 76 per cent believe the UK has the domestic capability to deliver world-class AI for the NHS, and I agree with this wholeheartedly, although given my role at Aire Logic I won’t pretend to be a neutral party.
We are a British, employee-owned company that has spent nearly 20 years building the infrastructure the NHS runs on, so naturally we think has exceptional domestic capability. But this number came directly from NHS leaders, with a clear-eyed acknowledgement that international partners will still be needed in some cases.
I read this as experts in their field saying they want a genuine choice about where its technology comes from, and do not feel they currently get one.
The finding I would most like people to reflect on, though, is the least convenient one, because when we asked what poses the single biggest threat to AI adoption over the next five years, leaders did not point at the technology and they did not point at the money. They said skills.
Workforce capability came top by some distance, well ahead of the legacy systems and technical debt that usually dominate this conversation, and only 6 per cent rate their own organisation as well prepared to adopt AI responsibly and at scale.
There is no procuring your way out of a problem like that, because no supplier, British or otherwise, can sell an organisation the capacity to absorb what it has just bought.
Building that capacity is slow, patient, and unfashionable work about people rather than platforms, and it is precisely the sort of thing that gets skipped when there is pressure to release something quickly.
What emerges from these 50 responses, then, is a group of leaders who are ambitious about what AI could do, confident in what this country is capable of building, and quietly frustrated that the strategy above them and the capacity beneath them aren’t aligned to real world challenges and concerns.
96 per cent say national AI strategy does not reflect their priorities, and out of fifty senior people, exactly one felt fully aligned with it.
That gap is one we need to close, and quickly, and it is why we are publishing this rather than keeping it as an internal talking point.
We would rather have the argument out in the open at AireFest, with the people who actually have to make this work, than spend another year talking about diagnostics while the real opportunity sits in a spreadsheet somewhere, waiting for someone to notice it.
Find out more about AireFest at airelogic.com/lp/airefest
