East Kent pilots Sanome’s MEMORI infection-risk AI

East Kent Hospitals University NHS Foundation Trust says it is the first NHS trust to trial MEMORI, an AI clinical decision support tool built by UK health-tech firm Sanome. The trust announced the pilot on 7 July 2026, running on a single ward: Harvey ward at Kent and Canterbury Hospital.

The pitch to buyers of clinical AI is that this is a low-lift deployment. MEMORI sits inside the trust’s existing Sunrise electronic patient record and reads data that is already being collected, so there is no new platform, no new integration project, and, for now, no bill: the pilot is offered cost-free as part of a longer-term trust-Sanome partnership.

What’s actually new

MEMORI generates an individual daily infection-risk score for each patient by analysing routine inputs already in the record: blood tests, observations such as blood pressure and temperature, medications, and demographics. When a patient’s risk changes, it alerts clinicians so treatment can start earlier and complications such as sepsis are caught sooner. It runs alongside existing monitoring like NEWS2 rather than replacing it, and BBC South East reported the tool required no new data infrastructure.

Sanome founder Benedikt von Thuengen, who started the company after his father nearly died of sepsis in an emergency department waiting room, told the BBC that “85% of clinicians are using the tool every single day.” Treat that as a vendor claim about early pilot uptake, not an independently audited metric: it is his figure, not a trust-published or externally verified one. Ward manager Julie Jones said the tool has “incredible potential” for spotting infection signs early. The first patient entered into the system, per the trust and HTN reporting, was Victor Casambros, 38, of Folkestone, a stroke patient on the ward since April, with observations logged by student nursing associate Bobbie Buddle.

Two clarifications worth holding. This is a one-ward pilot, not a trust-wide rollout. And claims circulating elsewhere, such as regulatory licensing status or detecting infections “up to three days earlier,” did not appear in the trust announcement or verified coverage, so we are leaving them out.

What this means for NHS trusts weighing clinical AI

The interesting part is the deployment model, not the algorithm. Most clinical AI stalls on integration cost and change management. MEMORI’s overlay approach sidesteps both: it co-developed with trust clinicians (nurses, healthcare assistants, consultants, therapists) and reads the EPR that staff already fill in, which is why daily-use numbers can climb fast. For procurement teams, that reframes the question from “can we afford the platform” to “what do we owe the vendor when the free pilot ends.” A cost-free trial inside a longer-term partnership is a familiar land-and-expand motion, and the leverage shifts once the tool is woven into a ward’s daily workflow. As we noted in our look at how deep AI adoption already runs in the clinic, the tools that stick are the ones that disappear into existing routines, exactly the pattern seen with AI medical scribes.

What to watch

Three signals: whether the pilot expands beyond Harvey ward, whether the trust publishes its own outcome data on sepsis and deterioration rather than leaning on vendor usage stats, and what the commercial terms look like when the free period ends. Coverage ran 7 to 8 July 2026 via the trust, the KentOnline report, the BBC, and HTN. Until independent results land, the strongest verified claim remains the deployment story, not the accuracy story.