NHS to expand AI use through federated data platform

NHS England is planning to ramp up the use of AI tools through its federated data platform (FDP), including products to support theatre scheduling, triage and discharge.

In a leaked internal memo dated 31 March 2026, seen by Digital Health News, Ming Tang, interim chief digital and information officer at NHSE, said that the programme would “accelerate our AI roadmap to drive even more improvements”.

She said this would include the rollout of assisted theatre scheduling, the Demand Centre for triage, and AI-assisted discharge summaries.

“These tools aim to optimise patient scheduling for theatre operations, transform triage, and reduce clinician administrative burden while boosting discharge rates,” Tang said.

Since going live in March, the Demand Centre has supported North West London in triaging 26,544 referrals, including more than 6,900 manually uploaded referrals.

Tang added that the roadmap includes enabling access to large language models for local development at every NHS trust and integrated care board using the FDP.

“This AI roadmap includes the enablement of access to large language models for local development against the canonical data model,” she said.

The memo also states that the FDP exceeded its 2025-26 target for adoption among acute trusts, reaching 86%, compared with a target of 85%.

Tang said that trusts using the platform are “already realising measurable clinical and operational benefits”.

She added that these benefits had been delivered while the programme focused on rollout.

“This phase has been much like laying the foundations of a building,” Tang said.

“It has involved an immense amount of work that has often not been visible from the outside, but it is the essential prerequisite for everything that follows.”

The programme will now shift from rollout to adoption, Tang said, as NHSE aims to increase use of FDP products within existing sites.

Tang’s comments come amid ongoing scrutiny of the FDP’s supplier Palantir.

Speaking in a parliamentary debate on 16 April, Zubir Ahmed, health minister, said that the £330 million contract would be subject to value-for-money assessments and could be revisited if better-performing suppliers emerge.

In the memo, Tang acknowledged criticism of the programme but said that the focus would remain on delivery.

“While media and campaign groups continue to raise questions about our technology supplier, the programme is focused on the evidence,” she said.

Meanwhile, an article published in The BMJ has criticised data used to support claims about the impact of the FDP as “flawed”.

Internal NHS data seen by the journal suggest that the data show “no meaningful difference between FDP adopters and non-adopters” and suggested that the study’s conclusion “conflates correlation with causation”.

An article by Rhiannon Mihranian Osborne, published on 13 April, suggests that “the performance improvements attributed to the FDP would have occurred regardless, as the health service recovered from the pandemic”.

Louis Mosley, executive vice chair of Palantir UK, which hold the contract for the FDP, posted on X criticising the article as “technically accurate, engineered to mislead”.

In response, Kamran Abbasi, editor in chief at The BMJ told Digital Health News: “The research article that he refers to was published in BMJ Health & Care Informatics, another journal from BMJ Group, and the study’s limitations that question its basis for scaling up the FDP are clearly acknowledged.”

Lee Rickles, chief information officer at Humber Teaching NHS Foundation Trust, told Digital Health News: “The FDP is the right vision, but we must challenge the assumption that a top-down acceleration via Palantir is the only route to success.

“For the FDP to work for providers like Humber, it must remain flexible enough to integrate with existing local innovations rather than mandating a one-size-fits-all timeline.”