The time-saving impact of generative AI in the public sector
A new report by the Alan Turing Institute, Department for Transport, and the ONS suggests up to 40% of public sector time could be supported by generative AI. In education, that figure rises to nearly half. Even in healthcare, a sector where human judgement is paramount, a third of time could be AI-assisted.
The potential is compelling. But there’s a catch: none of this works without high-quality, trusted data. And right now, most public sector organisations still don’t treat data like the asset it is.
Fragmented data, fragmented outcomes
Generative AI, for all its potential, is only as effective as the data it draws from. If your systems are fragmented, your records outdated, or your information buried in PDFs and policy silos, AI will struggle. The algorithms may be capable – but the environment isn’t ready.
This is a hidden barrier few talk about: public bodies are sitting on vast volumes of data without truly understanding its value, fitness, or risk profile. The result? Missed opportunities, inefficient investments, and growing risks to public trust.
How can you prioritise where to apply generative AI if you don’t know which datasets are reliable? Or where the gaps are? Or how much effort and cost is needed to bring your data up to standard?
Data as infrastructure, not an afterthought
To realise AI’s promise of less admin, more frontline time, and smarter service, we need to reframe how we think about data.
That starts with understanding what data you have, what it’s worth, and what condition it is in. Just as we wouldn’t build new roads on crumbling foundations, we shouldn’t build AI capabilities without confidence in the data beneath them.
By valuing data as a strategic asset, public sector leaders can:
- Prioritise AI investment in the most valuable, actionable areas
- Mitigate risk by exposing data that is poor quality, overly sensitive, or misaligned
- Align AI efforts with real-world service delivery and citizen needs
- Create transparency and trust about where AI will (and won’t) be used
Not all work is equal. Neither is the data behind it
The Turing study makes clear that AI won’t benefit all public sector roles equally. High-potential areas like education, social care, and administrative functions stand to gain most. But these are also the very domains where data tends to be unstructured, inconsistent, and locked in legacy systems.
If we want to automate lesson planning to free up teachers or help social workers spend more time with families and less on forms, we need to start by investing in the quality and structure of the data that enables those activities.
A smarter, safer way forward
As the Turing Institute rightly puts it, AI adoption is as much about responsible enablement as technical implementation. That means more than publishing ethical frameworks. It means knowing what data you have, its value to outcomes, and whether it can be trusted.
The UK government is right to support innovation that helps services run “smarter, not harder.” But running smarter starts with treating data not as a by-product, but as a strategic enabler.
Because ultimately, AI won’t transform public service delivery. People will: by choosing to invest in the data that makes meaningful transformation possible.
At Anmut, we help data leaders do just this
Using our approach to valuing data, we help public sector organisations treat data as a strategic asset, ensuring they can invest in data capabilities with confidence. For example, we helped a major government body value its nationally important data assets, creating the intelligence to achieve £100 million per annum in efficiency savings, and a 20% saving in manual labour from improved systems.
If you’re ready to make smarter AI investments and unlock the full value of your data, book a free discovery call with our team.