The digital workplace for clinicians: reducing friction in care delivery

The 10 Year Health Plan sets out a clear direction. Moving from analogue to digital. Making better use of data. Shifting more care into the community. And using AI to support clinicians, patients, and the wider NHS workforce. 

The 2025 NHS Staff Survey, published in March 2026, makes uncomfortable reading on this point. Only 56.06% of staff said they have adequate materials, supplies and equipment to do their job, a figure that has been drifting down for several years. Just 46.51% said they can meet all the conflicting demands on their time at work, and only 32.82% said there are enough staff at their organisation to do their job properly. Staff advocacy, the willingness to recommend the NHS as a place to work, fell from 60.79% to 58.05% in a single year. 

None of this is caused solely by clunky IT. But fragmented systems, repeated logins, and duplicated data entry are a consistent thread in why clinicians describe their day as harder than it needs to be. A survey of NHS healthcare professionals commissioned by Heidi found 80% saying their administrative workload has increased in recent years, with clinicians spending an average of 1.3 hours a day on admin on top of a full clinical schedule. 

Identity and single sign-on.
The GOSH trial, which came up in our piece on why the intelligent hospital needs security by design, found a 23.5% increase in direct patient interaction time, largely because clinicians stopped losing minutes to repeated logins across systems. That is a security win and a productivity win from the same piece of infrastructure. 

Ambient voice technology.
Early results are consistent enough to take seriously. A pilot at St George’s Hospital saved clinicians an average of 47 minutes per shift, roughly equivalent to one extra patient consultation per staff member per day. At Oxford University Hospitals, 87% of pilot users said ambient voice tools reduced their admin time, with most reporting savings of one to ten minutes per patient encounter. 

Copilot and everyday productivity tools.
More than 500,000 NHS staff now have access to Microsoft 365 Copilot, following a trial that found workers saving an average of two days of admin time a month. That is not a headline AI story. It is drafting letters faster, summarising handover notes, and cutting the time spent on routine correspondence. 

Workflow-native EPR design.
Maidstone and Tunbridge Wells NHS Foundation Trust cut the time spent recording clinical observations by 2.5 minutes per observation, a 50% saving, by configuring their EPR around how clinicians work rather than forcing clinicians to adapt to the system. 

It’s worth noting that the regulatory picture itself was evolving over this period. In April 2025, NHS England and the MHRA set out classification rules for ambient voice technology requiring any tool that went beyond simple transcription into summarisation to register as a medical device. That was a reasonable starting point given the sensitivity of clinical data, but in practice it slowed how quickly sanctioned AVT tools could reach clinical teams, which may have added to the pressure staff felt to find their own workarounds. In July 2026, the MHRA published updated guidance, aligned with NHS England, clarifying that AVTs used solely for transcription, summarisation, or drafting clinical correspondence don’t trigger medical device classification. That clarity should make it easier for trusts to sanction ambient voice tools going forward. 

That is not a reason to restrict access more broadly. It is a reason to get sanctioned, well-governed tools into clinicians’ hands quickly, using the kind of questions worth putting to AI suppliers before rollout, because the alternative, as the last eighteen months have shown, is unmanaged AI use continuing regardless. 

Passwordless and single sign-on through Entra ID, so clinicians authenticate once rather than repeatedly across the working day. 

Sanctioned ambient voice and Copilot tools, deployed with clear information governance, so staff aren’t left choosing between unmanaged consumer AI and no support at all. 

Device and endpoint management that matches how care is delivered, including virtual desktops for staff who move between wards, clinics, and community settings. 

EPR and workflow configuration built around clinical teams’ actual working patterns, not a generic template, following the same logic that delivered results at Maidstone and Tunbridge Wells. 

Flexible, mobile-first access, in line with the NHS’s Think Flex First ambitions, so digital tools support flexible working rather than tying staff to a single desk or device.

For many NHS trusts, the challenge is no longer whether digital transformation is necessary. It is how to make all the different pieces work together. 

AI tools need access to data. Data needs to move between systems. Systems need to integrate with existing EPRs and infrastructure. Clinicians need secure, simple ways to access the information they need. And all of this needs to happen within the NHS’s requirements for data protection, clinical safety, information governance, and cyber security. 

A trust does not necessarily need to implement every new technology at once. It needs to understand where its existing foundations are strong, where the gaps are, and which investments will create the greatest value. 

As we explored in Why Microsoft Purview is the first step to successful Copilot adoption in the NHS, AI adoption can expose existing problems with data access, permissions, and information governance. Getting those foundations right is part of being ready for AI in the first place. 

If you’re looking to reduce friction for your clinical teams, whether that’s identity and access, ambient voice and Copilot rollout, or device and endpoint strategy, we would welcome the conversation.