Why the UK Government Must Harness Private-Sector AI Innovation

Valued at more than £72 billion, the UK is the world’s third-largest artificial intelligence (AI) market, and a global hub for emerging innovation. Government adoption has accelerated rapidly, with more than £3.35 billion spent on AI contracts, infrastructure, and services since 2018, with the number of contracts increasing year on year.

The World Economic Forum’s Future of Jobs Report found that broadening digital skills will be the single most impactful trend reshaping organisations by 2030, with 60 percent of employers expecting this shift to transform their operations. For the public sector, this presents a significant opportunity: government departments that invest early in upskilling will be far better positioned to convert short-term technological disruption into long-term economic value.

The British government has already laid some foundations through the AI Playbook for civil servants and the creation of AI Growth Zones to build nationwide AI-ready infrastructure. However, guidance alone can’t close the capability gap. Government departments increasingly face a strategic question: should they build in-house, or should they draw on proven private-sector innovation to deliver AI safely, at scale, and with measurable impact?

AI Abandonment, and the Need to Prove ROI

A recent report by MIT found that 95 percent of enterprise generative AI pilots are failing, creating what the researchers term a “GenAI Divide” between the five percent of projects that deliver return on investment (ROI) and the vast majority that do not. Similarly, Gartner predicted that 30 percent of generative AI initiatives will be abandoned after the proof-of-concept phase by the end of this year. Therefore, going it alone without guidance into what is fast-moving landscape presents a considerable risk for a public sector that won’t have endless budgets for AI procurement.

The Right Honourable… Bot?

However, if done right, there are vast economic benefits. The Department for Science, Innovation and Technology (DSIT) believes AI efficiencies could save the UK government more than £45 billion a year. Some ministers have argued that these savings could extend to citizen-facing communication, such as using tools like ‘Caddy’, a customer service co-pilot that is being trialled by Citizens Advice.

Caddy was created by the Incubator for Artificial Intelligence (i.AI), an in-house DSIT team tasked with accelerating the responsible use of AI across government. To date, however, its projects have been relatively small in scale. Caddy is part of ‘Humphrey’: a wider productivity toolkit for civil servants, which includes Minute, a secure transcription service that made headlines as the first time AI was used in a meeting chaired by the UK Prime Minister, as well as Redbox, an AI chat and document tool. It was recently announced that Redbox is being discontinued since equivalent tools exist, and outperform its capabilities, in the private sector.

Learning From the Private Sector

Private sector AI innovation is driven by fierce competition and broad commercial demand for effective solutions. Tools evolve quickly and benefit from continuous improvement as AI vendors continue to leapfrog one other. In contrast, in-house tools like Caddy risk becoming outdated ‘oxbow lakes’ unless significant time and resources are assigned to maintain and develop them, not least because of the niche use-cases they represent.

i.AI’s pilots show that the government is willing to engage with AI-enabled services, and that citizens could benefit from faster, more-responsive departments. However, trust remains non-negotiable, and AI hallucinations can have catastrophic consequences in sensitive situations. With so much at stake, steps should be taken to improve accuracy through using Retrieval-Augmented Generation (RAG) methodologies that run AI output through verified foundation models, grounding responses against known government data. That way, government services can deliver the speed and convenience people expect without compromising trust.

Partnerships with private-sector specialists are already delivering substantial improvements. For example, the Driver and Vehicle Licensing Agency (DVLA) partnered with third-party providers to enhance citizen contact. As a result, the DVLA was rated Britain’s most improved public sector organisation in the UK Customer Satisfaction Index (UKCSI), the only public sector organisation to increase its customer satisfaction, and was the country’s fourth most improved customer experience (CX) provider overall. The DVLA’s chatbot now handles over 300,000 AI-powered web chats each month. Such AI solutions demonstrate what’s possible when proven private-sector innovation is applied to public-sector challenges: citizens will receive faster outcomes while developing trust in conversational services, and the government can free up human resources to better support those with more complex needs.

Building a Path Forward

The most successful public-sector transformations demonstrate that the considered implementation of AI can lead to drastic improvements at scale. The UK is well-positioned to become a global leader in the responsible use of AI within government, but only if it can harness the innovation already thriving across its AI economy, rather than trying to replicate it.

The UK can build better public services by combining the government’s commitment to ethical, secure deployment with the private sector’s ability to deploy solutions rapidly. Only through this partnership approach can government AI live up to its potential: empowering public servants and improving outcomes for citizens.