The UK government is investing over £100 million in a National Data Library and has published new guidelines for making datasets AI-ready. If your organisation holds legacy data – on tape, on paper, in obsolete digital formats – these developments could mark a step change in what that data is worth and what it will take to unlock it.
Most organisations have data they know is valuable but can’t easily use. It might be decades of seismic survey tapes sitting in a warehouse, patient records stored in paper envelopes, research archives on formats that modern systems cannot read, or heritage collections too fragile to handle. The data exists. The challenge is making it accessible, structured, and ready to work with.
That challenge becomes more urgent as the months roll on. In January 2026, the UK government published a progress update on its National Data Library (NDL), a flagship initiative backed by over £100 million as part of a broader £1.9 billion investment in digital priorities. Alongside it came new guidelines for making government datasets ready for AI. Together, these signal a step change in how public sector data is handled and viewed as a strategic asset. Expectations around how it’s managed, governed, and made usable are rising rapidly.
What the National Data Library means in practice
The NDL isn’t a single database. It’s a set of principles, infrastructure, and initiatives designed to make the UK’s vast public-sector data holdings more discoverable, accessible, and usable, particularly for AI.
The government has completed an extensive discovery phase, established an Expert Advisory Group, and is launching five ‘kickstarter’ projects testing new approaches to data access in areas from energy bill targeting to adult social care and AI-powered legal guidance for small businesses. It builds on initiatives already delivering real results. For example, the National Underground Asset Register generates over £400 million in economic growth each year by enabling secure data sharing across more than 600 organisations. The Health Data Research Service, backed by up to £600 million, will provide a secure single access point to national-scale health datasets by 2030.
The message from the government is clear: organisations that hold significant data assets need to consider how to make them usable at scale, not just store them safely.
New standards for AI-ready data
Alongside the NDL update, the government published guidelines and best practices for making datasets AI-ready. These matter because they define, for the first time, what ‘AI-ready’ actually means in a public sector context, and those standards will inevitably shape expectations more broadly.
The guidelines cite several assets and frameworks that aim to establish “the characteristics of whether available data is suitable for AI capabilities,” including a framework for AI-ready data, developed by the Open Data Institute (ODI).
According to the guidelines, a dataset can be considered AI-ready when it addresses:
- Technical optimisation – data is structured and formatted for machine learning systems.
- Data and metadata quality – data is accurate, complete, consistent, and well-documented.
- Organisational and infrastructure context – data governance, resources, and working practices are in place.
- Legal, security, and ethical compliance.
This isn’t just about having data in digital format; as the guidelines state, “the availability or openness of a dataset does not equate to AI readiness.” Many datasets remain, in the government’s own assessment, “fundamentally unfit” for AI use without significant preparation.
What this looks like on the ground – millions of NHS patient records
One of the most tangible examples is the NHS’s program to digitise the Lloyd George records (the paper-based patient files that have been a feature of general practice since the early twentieth century). Tens of millions of these records exist across England, many stored in overstuffed filing cabinets and basement cupboards.
NHS England has been digitising paper records and storing them in a National Document Repository (NDR), giving clinicians and practice staff digital access to historical patient information. The work is procured through the Digital Documents Solutions (DDS) framework, a £5 billion procurement vehicle run by NHS London Procurement Partnership with 46 approved suppliers.
The current DDS framework expires on 30 September 2026. When its successor is designed, the new AI-ready data guidelines are likely to shape the requirements. Suppliers and the organisations commissioning the work will increasingly need to ensure that digitised data meets standards for metadata quality, provenance, interoperability, and governance, the kind of standards the government’s guidelines now set out.
The Lloyd George program is a striking illustration of a much wider pattern. Across every sector, organisations hold vast quantities of legacy data with latent value, but only if it can be recovered, structured, and made ready for the systems and standards now emerging.
What this means if your organisation holds legacy data
Whether you’re sitting on decades of survey tapes, research archives on obsolete data formats, paper collections that need digitising, or large volumes of data locked in systems that no longer serve you well, the ground is shifting beneath you. The direction of travel, from “do we have data?” to “can this data responsibly sustain AI use across its full lifecycle?” is a question every data-holding organisation will need to answer.
Data standards are rising. More than ever, data needs to be accurate, complete, consistent, and enriched with metadata so it can be trusted and understood by both humans and machines. That applies whether you’re a government department, an energy company, a research institution, or a cultural heritage organisation.
Datasets are expected to have named owners, quality action plans, and clear legal bases. If your data is sitting untouched in storage, those expectations may feel distant, but the gap between “archived” and “usable” is exactly where value is being lost.
Format matters more than ever with the latest guidance stressing the importance of standard, widely compatible formats, application programming interfaces (APIs), and machine-readable documentation. Data locked in proprietary or obsolete formats, whether that’s a 40-year-old tape cartridge or a legacy digital system, will have diminishing value in a world where AI systems need to ingest, process, and link data at scale.
From recovery to readiness
At Ovation Data, this is the work we do every day. With 45 years’ experience, over 2,500 clients worldwide, and the ability to handle 275 magnetic and disk formats across more than 1,500 legacy tape drives and machines, we recover data from media that modern systems can no longer read. And we don’t stop at recovery.
We standardise and structure recovered data to comply with global standards, delivering it in open, widely compatible formats with no lock-in, no hidden fees, and no proprietary dependencies. Whether the starting point is sand-encrusted oilfield tapes, fragile century-old manuscripts, or petabytes of research data on obsolete cartridges, we prepare it for use in your specific environment. We ensure it’s ready for the decisions, discoveries, and AI applications that give it value.
We work as a single accountable partner across the full data journey: recovery, digitisation, structuring, secure storage, and access. Our data centres and cloud environments in both the UK and Houston meet strict security and compliance standards, and our systems are built to AWS Select Tier standards, giving you the flexibility to choose how and where you use your data.
The organisations we work with understand that legacy data isn’t a burden. It’s an untapped asset. The UK government’s National Data Library and AI-ready data agenda are making that case at a national level. The question is whether your data is ready to be part of it.
Sources and further reading:
National Data Library: progress update, January 2026 — GOV.UK
Guidelines and best practices for making government datasets ready for AI — GOV.UK
A framework for AI-ready data — Open Data Institute
National Document Repository: access and store digital patient documents — NHS England
Digital Documents Solutions framework — NHS London Procurement Partnership