Government adopts ‘test and learn’ strategy for AI and digital spending

The Government has announced a major overhaul of how AI and digital projects are funded in the public sector, aiming to cut waste, streamline innovation, and improve services.
The reforms, introduced by the Department for Science, Innovation and Technology (DSIT), will simplify complex funding processes and introduce a more agile, experimental approach to investment in AI and digital technology.
The changes, spearheaded by Technology Secretary Peter Kyle and Chief Secretary to the Treasury Darren Jones, will embed a ‘start-up mindset’ within government spending. The new approach will allow small-scale AI experiments to receive initial funding and grow incrementally based on demonstrable success, reducing the risk of large-scale budget misallocations.
The key reforms include:
- Staged funding for AI projects: Building on the success of the government’s experimental generative AI chatbot, GOV.UK Chat, a new funding model will allow projects to start small and scale up based on proven results.
- Faster approval processes: Current funding frameworks, which have been criticised for delaying decision-making and hindering innovation, will be restructured to ensure quicker development and deployment of digital solutions.
- Enhanced oversight and evaluation: The introduction of outcome-based metrics will ensure taxpayer funds deliver measurable improvements in public services.
The Government’s Plan for Change outlines ambitions to use technology to transform public services. In healthcare, AI-driven scheduling and diagnostic tools will help the NHS reduce wait times. Police forces will gain access to more reliable and efficient digital systems, enhancing their ability to tackle crime. Businesses will also benefit from streamlined digital processes, cutting bureaucracy and allowing them to focus on growth.
“Technology has immense potential to build public services that work for citizens. But a decades-old process has encouraged short-sighted thinking and outdated tech while stopping crucial innovation before it even gets going,” said Technology Secretary Peter Kyle.
“These changes we’re making ensure innovation is the default. We will help give AI innovators in government the freedom they need to chase an exciting idea and build prototypes almost immediately.”
The reforms follow a review which found that one in four digital systems used by central government is outdated, with some departments relying on systems where up to 70% of their technology is obsolete. These legacy systems come with high maintenance costs, often costing three to four times more than updating or replacing them, and many have been red-rated for reliability and security risks.
With the introduction of a ‘test and learn’ approach, the Government aims to drive greater efficiency in public sector digital projects, ensuring that new AI-driven solutions deliver real benefits while reducing unnecessary costs. The reforms mark a significant shift towards a more agile, outcome-focused strategy for AI and digital innovation in government.
