Data Foundations for AI in 2027: Testable Predictions for UK and Irish Organisations
Around a third of UK businesses now use artificial intelligence, but only one in ten of those uses it extensively. This outlook argues that the state of organisations’ records is one reason. Unstructured, legacy and paper records limit how far AI can move from pilot to production, and how well its use can be audited. Official surveys rank unclear use cases, cost and skills as bigger barriers, and the paper says so. It sets out five predictions for the end of 2027, drawing on data from the Office for National Statistics and the Central Statistics Office in Ireland, a BMJ investigation into paper records in National Health Service trusts, a Bank of England and Financial Conduct Authority survey, audit reports, and current UK, Irish and EU regulation. Each prediction names its data source, its threshold and the result that would prove it wrong. The coding rules for one analysis of public sector tender notices are registered in advance. A scored review will follow by December 2028.
Authors
- Damien Andrews (ORCID: https://orcid.org/0009-0001-9460-2365)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-10-29
- DOI
- https://doi.org/10.5281/zenodo.23768050
- Primary Topic
- E-Government and Public Services
- Type
- article
- Field-Weighted Citation Impact
- 0.00