NEUVIOR
NEUVIOR®
From pilot theatre to financial grip: how NHS CFOs should evaluate AI in medicines optimisation
Back to Blog
Finance

From pilot theatre to financial grip: how NHS CFOs should evaluate AI in medicines optimisation

NHS finance leaders do not need another AI pitch. They need a disciplined way to decide whether a technology can release value within the financial rules they are already being held to.

March 18, 20266 min read

From pilot theatre to financial grip: how NHS CFOs should evaluate AI in medicines optimisation

Target outlet(s)Primary: HFMA / Healthcare Finance. Secondary: HSJ Comment (only with an NHS finance co-author).
Primary audienceICB CFOs, trust CFOs, deputy CFOs, finance directors, chief pharmacists, digital sponsors.
Commercial purposeOpens PHARMORIS Commissioning and Provider conversations by framing medicines AI as a finance-grade decision tool, not a generic dashboard.
Draft length1006 words

Pitch to Healthcare Finance

NHS finance leaders do not need another AI pitch. They need a disciplined way to decide whether a technology can release value within the financial rules they are already being held to. That matters because the NHS finance environment has become less forgiving, not more. The 2025/26 planning cycle set a 7.1% efficiency requirement across the service, while NHS England’s 2026/27 finance business rules require both integrated care boards and NHS trusts to maintain breakeven positions and submit organisation-level plans. In that environment, any technology proposal that cannot survive finance scrutiny from day one is not innovation; it is noise. Medicines value is one of the few areas where this discipline can be applied quickly. NHSBSA’s latest official statistics show that NHS commissioners in England spent an estimated £20.9 billion on medicines, appliances and medical devices in 2024/25 after central rebates. Community prescribing alone accounted for 1.26 billion prescription items at a cost of £11.2 billion. When the spend base is that large, a low-basis-point improvement is material. But that does not mean every analytics or AI tool deserves a place in the plan. The wrong question is, “How sophisticated is the algorithm?” The right question is, “Can this product help us make, approve and deliver better medicines decisions quickly enough to matter this year?” There are seven tests every CFO should apply.

1

The first is whether the product is tied to a defined financial problem. “Improving medicines optimisation” is

too vague. “Reducing premium-priced generic leakage”, “accelerating best-value biological uptake”, “improving contract compliance” or “shortening time from signal to approved action” are real problems. The more specific the problem definition, the easier it becomes to assign ownership, measure progress and stop wishful savings from creeping into the case.

2

The second test is whether the baseline is credible. Public data is useful, but it is not magic. NHS England’s

medicines optimisation resources and NHSBSA dashboards already give systems access to comparators, prescribing trends and opportunity signals. The NHSBSA itself explicitly states that the metrics in the national medicines optimisation opportunities dashboard are only a starting point and that additional national and local data, as well as more sophisticated comparisons and local knowledge, are often essential to interpret the data properly. A serious vendor should therefore be able to show where the baseline comes from, what the time window is, what assumptions are being made and what the limitations are.

3

The third test is whether value is linked to deployment, not just detection. This is where many digital

propositions fail. A chart that identifies variation is not yet a saving. An alert that shows a price anomaly is not yet a saving. A ranked opportunity only becomes financially meaningful when it is connected to an evidence pack, a decision path, a local owner, a timeline, an adoption curve and a realised-value review. If the operating model stops at “insight”, the CFO should assume the benefit case is overstated.

4

The fourth test is assurance readiness. NHS England’s Digital Technology Assessment Criteria now uses an

updated form from February 2026, with the previous form retired from 6 April 2026. DTAC remains the national baseline for digital health technologies and covers clinical safety, data protection, technical security, interoperability, and usability and accessibility. NHS England is also clear that products should be assessed against DTAC regardless of procurement route. That means finance leaders should not treat assurance as someone else’s problem. If the supplier cannot show a coherent DTAC posture, clear data protection documentation and an honest statement of what approvals do and do not yet exist, procurement drag is almost guaranteed.

5

The fifth test is whether the claims are bounded. NHS data is not universally live. Secondary Care Medicines

Data is monthly. Some benchmark measures are contextual rather than prescriptive. Some dashboards do not capture clinical indication. Even the new SGLT2i opportunities dashboard notes that not every patient shown will be suitable for switching and that clinical judgement remains essential. A credible business case therefore needs ranges, caveats and confidence levels. Over-precise savings numbers are usually a warning sign, not a comfort.

6

The sixth test is implementation burden. NHS England’s Buyer’s Guide to AI in Health and Care remains useful

because it forces buyers to move beyond the demo and ask practical questions: what problem is being solved, do the performance claims hold up, will the product work in practice, will staff support it, and how will the contractual and procurement arrangements stand up over time? Finance leaders should adapt that logic ruthlessly. If the tool creates another meeting series, another spreadsheet workflow or another disconnected governance pack, it is probably adding cost before it adds value.

7

The seventh and final test is how value will be reviewed in public. Under the 2026/27 finance rules,

organisations are being pushed toward clearer public reporting of financial performance and greater transparency around plan delivery. So the review mechanism matters. Finance teams should ask for a simple structure: forecast value, approved actions, delivered value, current blockers, and variance to plan. If the vendor cannot explain how the organisation will see forecast versus realised benefit after ninety days, the proposition is not board-ready. This is why the best digital purchases in medicines are rarely technology-only decisions. They are joint operating decisions. The sponsor group should usually include the CFO or deputy CFO, the chief pharmacist or medicines optimisation lead, and the digital or data lead. Finance defines the standard of proof. Pharmacy defines clinical appropriateness. Digital defines assurance and implementation fit. If one of those functions is missing, the probability of either delay or disappointment rises sharply. None of this is anti-innovation. It is the opposite. The NHS does not need fewer ideas; it needs fewer undisciplined buying decisions. In a tight financial regime, a technology should be expected to help the organisation do one thing better than the manual alternative: identify the right opportunity, justify it properly, deploy it with less friction, and show what value actually landed. That is the standard medicines AI should now be held to. Not pilot theatre. Financial grip. Research base used in drafting

Research Base Used in Drafting

  • NHS England - 2025/26 priorities and operational planning guidance (30 January 2025; updated 22 August 2025).
  • NHS England - Financial performance update / NHS England Board minutes on 7.1% efficiency requirement and 2025/26

planning position (March-May 2025).

Research Base Used in Drafting

  • NHS England - NHS finance business rules from 2026/27: guidance for integrated care boards and NHS trusts (16

December 2025).

Research Base Used in Drafting

  • NHSBSA - Prescribing Costs in Hospitals and the Community - England 2024/25 (20 November 2025).
  • NHSBSA - Prescription Cost Analysis - England 2024/25 (5 June 2025).
  • NHS England Transformation Directorate - DTAC guidance for buyers and suppliers (updated February 2026).
  • NHSX / NHS England - A Buyer's Guide to AI in Health and Care.
NEUVIOR®

Building pharmaceutical Infrastructure with intelligence and integrity.

Infrastructure
Intelligence
Integrity