Somewhere in the corridors of St George's, Epsom and St Helier, a clinician is probably still walking to find a signal. That image, unglamorous and entirely ordinary, sits at the heart of the group's newly published digital strategy for 2026 to 2030, a document that asks for £23 million to £25 million simply to harden networks and extend wifi coverage across its estate. It is a modest, almost apologetic request, buried in a plan that elsewhere promises ambient voice technology, Microsoft Copilot and an AI centre of excellence. The strategy calls its existing technology landscape fragmented and ageing. Few phrases in NHS digital policy carry more weight, because they describe a condition shared well beyond south-west London.

This is the quiet contradiction running through much of the health service's digital agenda. National ambition has moved decisively towards artificial intelligence, automation and a single patient front door through the NHS App, all wrapped in language about productivity and modernisation. Meanwhile many trusts are still negotiating with connectivity that would not surprise anyone working in a mid-sized office a decade ago. The St George's group is unusually candid about this, setting out its intention to rationalise systems, optimise its electronic patient record and build a group-wide data capability at the same time as it commits to frontier tools. Reading the two ambitions side by side, remediation and reinvention, it becomes hard to avoid the sense that NHS strategy documents are increasingly asked to do two jobs that pull in opposite directions.

The political backdrop makes this harder to ignore rather than easier. NHS England's absorption into the Department of Health and Social Care is still working its way through Parliament, with the legislation at committee stage and full abolition not expected before April 2027. Staff across NHS England and DHSC have spent much of the past year managing consultations, redundancy applications and organisational redesign rather than delivering the digital infrastructure programmes those same reforms depend upon. Yvette Cooper's arrival at the department, following a rapid succession of health secretaries this year, brings a fresh mandate focused heavily on maternity services and social care. Digital infrastructure has not disappeared from the agenda, but it now competes for ministerial attention with a reorganisation that has absorbed enormous institutional energy without yet delivering the promised simplification.

Set against that, NHS England's own sepsis modern service framework is instructive. It lays out a ten year digital roadmap spanning the single patient record, the Federated Data Platform, machine learning and wearables, aiming to cut sepsis deaths and complications by a quarter by 2035. The ambition is entirely reasonable. What is less clear is whether the infrastructure beneath it, the same infrastructure St George's is asking £23 million to shore up, can bear that weight consistently across more than two hundred acute trusts, many of which have not published anything as candid as this group's admission of fragility.

None of this means digital transformation should wait until every network cable is replaced. Patients benefit from the NHS App, from faster referrals, from data that follows them between services rather than sitting in silos. But there is a real risk in allowing national strategy to talk primarily about AI adoption while trust-level reality remains preoccupied with basic resilience. That mismatch has consequences for safety, since ambient voice tools and automated processes depend on networks that do not fail, and consequences for public trust, since patients asked to manage their care through an app have little patience for services that cannot yet keep a hospital's wifi running reliably.

The St George's strategy deserves credit for naming its own limitations rather than papering over them with promises of artificial intelligence. Other trusts, and the department now absorbing NHS England's functions, would do well to follow that example. Modernisation built on foundations still being repaired is not transformation. It is exposure, dressed in more confident language than the underlying infrastructure can support.