At NHS ConfedExpo in Manchester last month, a senior digital leader stood in front of a room of health service executives and asked, only half rhetorically, how many of the workforce were still barred from using large language models at work. A few hands went up in disagreement. The exchange captured a tension that now defines much of NHS technology policy: an institution simultaneously racing toward automation and struggling to translate enthusiasm into governed, scaled deployment.

That tension is not unique to Britain. A new report from KLAS Research, drawing on interviews with 213 individuals across 182 healthcare organisations in 43 countries and territories outside the United States, finds that artificial intelligence has for the first time become the top-cited technology investment priority in every region measured. Fifty-seven per cent of respondents named AI among their top three priorities for the coming two to three years, a jump of fifteen percentage points on the prior year. Every other category of investment held steady or slipped.

The reasons will be familiar to anyone working in the NHS. Staffing shortfalls remain acute across nursing, primary care and allied health professions in most developed systems, and administrative burden continues to consume clinical time that could otherwise go to patients. Budgets, meanwhile, are tightening almost everywhere, forcing health systems to look for productivity gains that do not depend on hiring more people. An ageing population is adding to demand at a pace that workforce growth cannot match. These are the same pressures that have driven NHS England to open a national supplier registry for ambient voice technology, published in January, requiring vendors to hold Class 1 medical device accreditation and a current assessment under the Digital Technology Assessment Criteria.

What the KLAS data adds is scale and comparison. The report notes that AI investment is reshaping adjacent spending, with capital moving away from generic infrastructure projects and toward cloud migration, data modernisation and cybersecurity frameworks built to support AI workloads. Seventy-three per cent of international health networks are now actively deploying clinical workloads on public or hybrid cloud platforms, a marked shift from planning to execution. Health systems are also turning increasingly to consulting firms to implement these deployments, an acknowledgement that internal digital teams are rarely equipped to manage projects of this complexity alone.

Ambient clinical voice technology stands out as the area generating the most enthusiasm, according to the report, a finding that maps closely onto the NHS's own experience. A Great Ormond Street-led study found ambient notetaking freed clinicians to spend nearly a quarter more time with patients, and a pilot at St George's reported savings equivalent to one additional consultation per clinician per shift. Yet the same NHS leaders celebrating those figures have also warned against what they call "pilotitis": a pattern of repeated small-scale trials that never graduate into system-wide, governed deployment.

That warning is where the global data becomes most instructive for NHS leadership. The KLAS report is explicit that most organisations remain in an early, foundational phase, with over a third focused on governance and readiness rather than procurement. The sequencing question, in other words, precedes the purchasing question. For NHS trusts weighing further AVT contracts, or for NHS England as it considers the next phase of its ambient scribing registry, the global pattern suggests that governance capacity and data infrastructure will determine which investments actually deliver, rather than the scale of ambition behind them. Life sciences and health-tech suppliers eyeing the NHS market would do well to note the same caveat: enthusiasm for AI is now close to universal among health system buyers, but the organisations able to convert that enthusiasm into working systems remain a minority.