Moving beyond personal productivity to transform the entire NHS. Taking control by building a new operation model.
Sponsored and written by
The NHS has finished building its digital foundations. The £1.9bn electronic patient record rollout is done. Most acute trusts now have a system. That was the main game for the better part of a decade.
Despite genuine progress, the EPR investment has not yet delivered its productivity promise. According to NHS England’s own usability survey, only 34 per cent of staff say their EPR makes them more efficient. The paper is largely still there, too.
The 2026/27 NHS financial outlook is relatively stable, with only 10 providers forecasting deficits totalling £420m – down from £2.2bn previously. However, this improvement stems from severe cuts: two-thirds of trusts are reducing clinical staff, 90 per cent are cutting admin roles, and more than half of NHS workers report working while unwell. While finances stabilise, workforce conditions deteriorate.
The government further introduced a 10-Year Health Plan that proposes AI as a practical tool for every clinician, projecting that technology could release £13bn worth of time. There is serious money behind it – a Frontline Productivity programme likely to attract investment at EPR scale – and the expectation to double down on AI in the soon-to-be-released workforce plan. But the NHS has been here before: large commitments, genuine intent, then a gap between what was promised and what landed.
The personal productivity gamble
Most NHS AI investment has gone into tools designed to help individuals work faster. Ambient voice documentation. Individual Copilot technology. AI scribing. These are real tools with genuine use cases – early evidence suggests ambient voice returns more than five minutes per consultation and meaningfully improves documentation quality. For the clinicians who get value from them, that value is real.
The NHS’s cost and capacity challenge is not a personal productivity problem. It is an enterprise productivity problem. Think how work moves across departments, how decisions get made, how hundreds of manual handoffs between systems and teams accumulate into structural inefficiency. Personal productivity tools don’t touch that. They optimise the edges while leaving the operating model intact.
This is the question hanging over ambient voice and Copilot as instruments of system-level change: not whether they help individual clinicians – they do – but whether per-consultation time savings ever aggregate into whole-system productivity gains the NHS needs and can ultimately take to the bank.
The data the NHS doesn’t talk about enough
NHSE’s own analysis of people services makes the scale of the untapped opportunity plain. Today, fewer than 5 per cent of HR processes in NHS trusts are automated. AI adoption is lower still, at just more than 1 per cent. More than half of people professionals sit between bands 2 to 5 – indicating there is still an industry of transactional work that needs to be resolved.
The assumption that the low-hanging fruit has already been picked is wrong. The orchard is largely untouched.
What enterprise productivity requires
Agentic Process Automation (APA) works differently from the tools that have dominated NHS AI investment so far. Rather than assisting an individual with a task, it operates across an entire service – connecting systems, executing decisions, managing handoffs, running processes end-to-end without requiring a human to push work from one stage to the next.
The NHS’s efficiency problem lives in the steps between systems, and in the work that falls between job descriptions. For example, in a recruitment process that takes months, it is not because any individual is slow, but because the process itself has dozens of handoffs, each requiring someone to do something manual before the next person can start. APA restructures that process and removes the handoffs so people can deliver better outcomes.
The evidence on this is already concrete. Kent Community Health reduced its temporary staff spend by 36 per cent in approximately six months via APA. Real numbers are being delivered by real trusts, landing in the same financial year they were invested in. That in-year, cashable return is the currency that gets attention from a chief finance officer – not three-year NPV projections, but savings a finance leader can account for now.
Leicester leads the way
University Hospitals of Leicester (UHL) is the most developed example of this approach in practice. UHL isn’t deploying automation to improve a workflow; it is rebuilding its people services operating model from the ground up, with APA as the enabling architecture, delivering a unified operating model with University Hospitals of Northamptonshire.
The ambition is to make 50-70 per cent of administrative work fully or partially autonomous. The method is grounded: each wave is designed to be cash-positive before the next begins. The programme has already identified more than £18m in cost reduction over three years and is projected to free more than 60,000 hours for frontline care.
“Healthcare is a team sport,” says Clare Teeny, chief people officer. “I could see a vision to create an agentic colleague, which would free up the capacity of the people within the team to do the work only they can do.”
UHL’s chief executive, Richard Mitchell, puts the ambition plainly: by 2028, he wants to walk the wards and hear from staff that their jobs have fundamentally changed.
That is the right frame of mind. Not AI as a threat to jobs, but AI as the thing that finally removes the work nobody wanted to be doing anyway.
Taking back control
NHS technology circles often view AI and automation as controversial and disruptive. Yet this same technology can restore control over processes, costs, and outcomes. When applied enterprise-wide to transactional work, automation delivers predictable, scalable results.
Some trusts are already generating real in-year returns through live programs that transform how work gets done. The opportunity is largely uncaptured despite proven tools and compelling financial cases that close within the year. The NHS doesn’t need to pioneer this – it can follow existing blueprints from successful trusts.
While the dominant AI narrative focuses on risk and deferred rewards, evidence from trusts running automation at scale shows tangible returns, immediate savings, and improved staff working conditions. Their success story deserves louder recognition.
