Bailie Review: Budget 2026 Bets £700m and an Independent Inquiry on NHS AI Productivity

Another Review. But This One Has Teeth.

The Autumn Budget, delivered on 28 October, contained a lot of AI promises. For the NHS, the headline was an independent review of productivity and artificial intelligence, commissioned jointly by Chancellor John Healey and Health Secretary Yvette Cooper. Its chair is Mark Bailie, chief executive of Compare the Market's parent company and an NHS England non-executive director. Findings are due in early 2027, timed deliberately to feed into the next spending review.

The remit is straightforward, if enormous: identify the practical opportunities AI and technology present to transform patient services and lift NHS productivity. In Healey's own words, the goal is to let staff "spend their time with patients and not on admin."

The Kings Fund noted, with barely concealed weariness, that "the NHS clearly doesn't have a productivity problem when it comes to politicians producing productivity reviews." That quip lands because it carries weight. Yet this review is structurally different from many of its predecessors. It is explicitly tied to spending-review allocations, meaning Bailie's recommendations could directly determine which AI programmes the NHS funds at scale from 2027 onward. That is leverage previous reviews rarely had.

What the Budget Actually Puts on the Table

The NHS-specific announcements sit within a wider AI package worth over £700 million. The key components:

• £140 million directed at the Sovereign AI Unit to invest in strategically important British AI firms
• £500 million for AI safety and security infrastructure
• £60 million AI Pathfinder Fund for cross-government barriers and opportunities
• £100 million to expand robotic-assisted surgery across England, backed by a national AI-powered surgical outcomes registry
• £190 million for the NHS in England to manage winter pressures in 2026-27
• A new British Data Centre Covenant, specifying tests that strategic data centres must meet, with health data sovereignty a named priority. Healey was blunt: "When the NHS uses AI to treat patients, we want that data here."

The £15 million AI Challenge Prize is the detail most likely to animate the healthtech startup community. It runs as an annual competition, rotating across public service priorities. The inaugural round targets the NHS, with a top award of up to £2.5 million for the most promising solution to a priority healthcare challenge. For early-stage medtech and diagnostics firms, that is a credible incentive, and the kind of competition-based commissioning that sidesteps the glacial pace of standard procurement.

The Adoption Gap Nobody Wants to Name

The review arrives at an inconvenient moment for NHS AI optimism. Around 90% of NHS AI tools remain stuck in pilot phases, according to NHS England's own assessment, a figure that has barely shifted despite years of acceleration rhetoric and significant public investment. The Budget documents note that AI diagnostic tools have already helped over four million patients access faster lung cancer diagnoses, and that the NHS App now serves more than 39 million users. Progress is real. But it is uneven and partial.

David Hill, Partner in Healthcare Regulation at Browne Jacobson, put it plainly: "Commissioning an independent review into the opportunities AI presents for NHS productivity is a sensible first step, but the emphasis needs to be on rigour as much as speed." That tension, between political urgency and clinical governance, is precisely what the Bailie Review will have to resolve.

The robotic surgery investment illustrates the complexity. The £100 million is drawn from the NHS Transformation Fund, and the accompanying national registry is designed to use AI to track outcomes and surface uneven access across trusts. It is sound in principle. But deploying a genuinely interoperable, AI-readable surgical registry across England's fragmented trust landscape is a non-trivial technical and governance problem. The Budget announces the ambition; the review will have to decide whether the infrastructure exists to support it.

Sovereignty, Data, and the Quiet Shift in Procurement Logic

The Data Centre Covenant is worth watching beyond its headline framing. By setting explicit tests that strategic data centres must meet, covering community investment, local job creation, and national data sovereignty, the government is signalling that future NHS AI contracts will carry sovereign data requirements as a baseline. That is a material change for any US-headquartered vendor currently holding or processing NHS patient data.

It also creates a new commercial opening for UK-domiciled AI infrastructure and life sciences technology firms. AIBD analysis of IPO trademark data shows 778 UK Class 5 trademark filings in 2026-Q4, a sharp fall of 68.6% against the prior period (AIBD analysis of IPO (TMD) data, as of 2026-11). That contraction may partly reflect the chill in early-stage pharmaceutical and diagnostics branding activity, but it also suggests the pipeline of new life sciences brands coming to market has thinned just as public procurement appetite for health AI is about to expand. The timing creates a gap that established players, rather than newcomers, may be best placed to exploit.

What Comes Next

The Bailie Review is expected to report by spring 2027. Between now and then, the NHS AI and productivity landscape will be shaped by the MHRA's expanding AI Airlock regulatory sandboxes, the government's acceptance earlier this year of all 44 recommendations from the National Commission into the Regulation of AI in Healthcare, and the practical rollout of ambient voice technology and AI triage tools already underway across trusts.

The RCN's general secretary Nicola Ranger was not persuaded the budget treated workforce as seriously as technology: she said health and care had been treated as an "afterthought" rather than a system whose people are the precondition for any AI gain. That criticism, too, falls within Bailie's remit.

So: a review with teeth, money with conditions, and a question that has been asked before. Whether the answer, this time, is different.