When Whitehall Buys Assurance by the Framework, the NHS Buys It by the Contract

The announcement landed with the flat procedural tone these things usually carry. HM Revenue and Customs has appointed Atos to lead governance and configuration across its low-code application ecosystem, a three-year arrangement worth up to £78.2m with an option to extend for two more. The head of Atos UK&I, Michael Herron, described it as an extension of a long-standing partnership. In Whitehall terms, it barely registers as news. HMRC has been buying digital modernisation in instalments for years, through a framework built precisely so that no single supplier, and no single failure, can hold the department's legacy estate hostage.

That framework is DALAS, the Digital and Legacy Application Services agreement Crown Commercial Service established to wean HMRC off ageing infrastructure. Since it came into effect, HMRC has awarded well over a billion pounds in contracts under its umbrella, spread across more than two dozen suppliers, with recent lots going to Netcompany, IBM and Accenture alongside Atos. The logic is straightforward. Legacy technology risk is diversified across a competitive panel, refreshed periodically, with governance and assurance functions bought as a distinct service layer sitting above delivery. If Atos underperforms on the low-code ecosystem, HMRC has both the contractual mechanism and the market of alternative suppliers to respond.

Set that against how the NHS has approached its own equivalent problem. Health data infrastructure sits overwhelmingly with a single supplier and a single contract of enormous political and operational weight. The Federated Data Platform, built and run by Palantir, was meant to knit together fragmented NHS trust systems and unlock efficiency gains across the service. Instead it has become a case study in the risks of concentration. The platform's projected benefits had been overstated by £1.7bn, and the latest figures from the National Infrastructure and Service Transformation Authority put the whole-life cost at £1.1bn against a forecast benefit of £808m. There is no framework of alternative suppliers standing behind that contract, no competitive panel to fall back on if delivery falters, and no separate assurance layer independent of the vendor itself marking its own progress.

This is not a case of HMRC getting technology right while the NHS gets it wrong. Framework-based procurement carries its own costs, in coordination overhead, in the difficulty of holding any single supplier fully accountable for outcomes that span several contracts, and in the slower, more bureaucratic pace such structures tend to produce. But it does distribute risk in a way that single-vendor mega-contracts do not, and that distinction matters more for the NHS than for almost any other part of the state. A tax authority that stumbles on a low-code deployment inconveniences taxpayers. A health data platform that overpromises and underdelivers affects how quickly hospitals can identify patients waiting longest, how effectively trusts can plan capacity, and ultimately how care gets prioritised.

The comparison sharpens further with what has followed Palantir's contract. Ambient voice technology is now being rolled out across clinical settings with regulatory oversight still catching up to the pace of deployment. Advanced Foundation Trusts are being given more latitude to run their own technology procurement, which will only deepen the fragmentation NHS leaders are meant to be solving. Each of these initiatives adds another vendor relationship, another point of dependency, without the kind of structural safeguard HMRC has built into DALAS almost by default.

None of this means the NHS should simply copy Whitehall's model wholesale. Health data carries sensitivities and integration demands that a tax department's application estate does not. But the underlying principle, that legacy technology reform should not rest on the performance of one contract with one supplier, is not a health-specific insight. It is basic risk management, and it is one HMRC has quietly practised for years while the NHS has repeatedly bet its digital future on arrangements too large to easily unwind. The Atos contract will not make headlines beyond the trade press. It is nonetheless a useful reminder of what disciplined procurement looks like, and how far the NHS still has to travel to get there.