Keir Starmer presses home AI plans for government

The UK Government has stepped up its mission to digitalise public services, with Prime Minister Sir Keir Starmer pressing home the cost savings related to deploying AI across government.

“There are up to £45 billion worth of savings and productivity benefits, ready to be realised,” he said.

The government said it has adopted the mantra, “No person’s substantive time should be spent on a task where digital or AI can do it better, quicker and to the same high quality and standard.”

However, Science and Technology Secretary Peter Kyle said it is “almost certain” civil service staff numbers “will go down” as AI is used in government departments to drive improvements.

In an interview with Sky News, Kyle said that “More than half or about half of all transactions is carried out by government or analogue. “So that is for example, the DVLA opening 45,000 envelopes every single day.

“HMRC is picking up the phone 100,000 times every day – this is not the way we should be doing government. This is not the way we should be running a country in the 2020s.”

DSIT this week said that initial tests of an AI helper for call centre workers, built in partnership with Citizens’ Advice, showed that it could halve the amount of time it takes call handlers to give responses to complex questions, from consumer rights to legal support.

The digitisation will also include a new apprenticeship scheme, TechTrack, which aims to bring 2,000 apprentices into public sector departments by 2030.

“There is a £45 billion jackpot to secure if we use technology properly across our public sector – but we can’t hope to come close to securing that if we don’t have the right technical talent with us in government,” said Kyle.

“Not only will these changes help fix our public services, but it will save taxpayer cash by slashing the need for thousands of expensive contractors and create opportunities across the country across the country as part of our Plan for change.”

Tech firms welcome AI push

Unsurprisingly, the tech industry – especially those developing AI systems – welcomed the announcement.

Damian Stirrett, group vice president & general manager UK & Ireland at ServiceNow, noted that AI-driven automation is already cutting error rates by up to 50 percent and reducing casework time by over a third in the private sector.

“AI, and Agentic AI in particular, will be central to a successful and sustainable transformation, and it is promising to hear Keir Starmer’s plans for greater use of AI in the civil service,” he said. “Agentic AI is the new frontier, delivering predictability and efficiency across organisations. Investing and implementing Agentic AI will enable the civil service, and public state as a whole, to meet demands faster, smarter, and at scale.” 

Ian Jeffs, UK&I country general manager at Lenovo Infrastructure Solutions Group, described the move as “a proactive step towards improving efficiency and modernising public services.”

However, he said that while investment and confidence in AI continue to grow, critical challenges remain around training staff, poor data quality and scaling initiatives. “Careful implementation is required by governmental departments to ensure the right frameworks are in place to enable innovation and productivity. For it to be considered a success, the UK Government must be proactive and implement impactful AI solutions that showcase ROI and turn words into actionable outcomes.”

“Government systems are complex, with a lot of important information being shared between teams and across disparate systems. There are some obvious quick wins for the Government in automating tasks through AI that are currently reliant on outdated processes and require a lot of manual input. This will free up civil servant time, but AI can only be as effective as the data that is available to it,” said Rupal Karia, GM – UKI & MEA at Celonis, which helps UK organisations such as the NHS cut costs through process mining technology. 

“However the risk is in deploying AI without systems speaking the same language and having the same organisational context.”

Karia said this is only possible through Process Intelligence platforms, and Celonis would build on its work with the Cabinet Office to streamline its Shared Services for Government computer system which serves half a million civil servants. 

“Without getting the processes and data in order first there is a risk of the government adding another layer of complexity to the civil service instead of easing workloads and saving costs.”

James Hall, VP and country manager UK&I at Snowflake said centralising, securing and scaling data for AI applications, enabling organisations to confidently achieve measurable outcomes, will be key to AI really delivering for working people.

“AI systems are only as powerful as the data they’re trained on, making high-quality, accessible data essential. Rather than struggling with a reactive approach tackling new technology case by case which can be costly, the Government should also focus on AI governance frameworks that proactively seek ways to address challenges, such as privacy and autonomy. Data sharing across trusted ecosystems will help the Government reach its goals, but also ensure its use of AI is accountable, free of bias and measurable.”

Creating AI literacy in government

Elsewhere, Neal Riley, co-founder and general manager of digital transformation specialist Salable, part of The Adaptavist Group, said the mandate for greater innovation and efficiency will not be a simple undertaking.

“Organisations that benefit the most from AI rollouts are often digital natives who already have a culture of adopting modern technology and trends. The Government historically doesn’t fit this kind of description, and a serious technology upgrade needs to be applied before this kind of technology can actually meet the full expectations that Starmer set out.”

Riley also pointed out that “the quantity of services provided that represent the underpinnings of the success of Starmer’s vision. 

“This can’t happen with a work force that is AI illiterate. In order for every government employee to be able to gain the most out of this technology, much like early digital transformation efforts before it, requires a basic understanding of the core tech, how and where it can be applied, and meeting skills gaps that exist today. This human enablement: ensuring that all employees who are seeking to gain these benefits have the understanding that they need, has to be a priority.”

_{This article has been updated}