Organisations like mine
68% of physicians say the NHS lacks the infrastructure to deploy AI, and integration is the barrier they name first
Royal College of Physicians, 2026-08-13
Responding to the GMC's Workplace experiences 2026 report, the RCP cited its own survey: 68% said the NHS lacks the digital infrastructure to deploy AI effectively, and 70% named poor integration between digital systems as the biggest implementation barrier. Senior Censor Dan Furmedge wants AI introduced cautiously, with governance.
TakeawayYour infrastructure argument now has a national citation from a royal college, not just your own risk log. Put the number in the next board paper.
rcp.ac.ukGloucestershire Health and Care is buying digital confidence in units of people: 500 this year, champions at 10% of the workforce
Gloucestershire Health and Care NHS FT, HTN, 2026-08-18
The strategy to 2031 commits to targeted training for 500 colleagues in 2026/27 and expands the digital champions network to include data and AI champions, targeting ten percent of the workforce. It promises to prioritise high-value AI use cases and embed clinical safety, ethics and accessibility in every digital change.
TakeawayA ten percent champion target is a number you can plan and defend. Eight trust strategies here in a fortnight, and this is the first to size the workforce side.
htn.co.ukGDS solved the "we cannot test on the real data" problem by generating fake government email first
Government Digital Service, Data in government, 2026-08-05
The AI Data Readiness Incubator and the Government KIM team built two tools. Tiger Heron generates synthetic emails from real ones so teams can experiment safely. Cuckoo clusters large volumes by content into an interactive view, so a records professional triages groups instead of messages. Phase one was deliberately exploratory.
TakeawayThe synthetic data step is the transferable one. It is how you build and test against a sensitive corpus before information governance has agreed anything.
dataingovernment.blog.gov.ukCoffee break
digital maturity, and how to measure it in people
Digital maturity is usually reported as an organisational score. The version that constrains an AI programme is the workforce one: whether the median member of staff can be handed a tool and get value from it without a project wrapped around them.
England has two instruments and they measure different things. The **Digital Maturity Assessment** is an annual self-assessment by every trust and ICB against What Good Looks Like. It is organisational: useful for benchmarking, weak as a guide to what one person can do. The **Digital Capabilities Framework** works at the level of the person, and its most practical finding is that self-reported **digital confidence** predicts actual capability well enough to serve as the primary indicator, and is far cheaper to collect.
So: baseline confidence by staff group before rollout, not after. Segment rather than report a trust average, because the average hides the groups that will not adopt. Champion coverage is your input measure, confidence movement your output measure.
