Leveraging AI in the UK public sector

April 21, 2025

Image: © pixdeluxe | iStock

Simon Dobbs, Director of Technology at Bytes Software Services, explains how to leverage artificial intelligence in the UK public sector effectively

The UK Government recently announced plans to put artificial intelligence (AI) at the forefront of their strategy and to put “government in your pocket”, come at an opportune time for UK PLC.

According to PWC’s 28th Global Annual CEO Study, AI is already at the heart of many CEOs’ strategies. The CEOs surveyed reported that “from GenAI: 56% reported efficiency gains, while one-third saw profitability (34%) and revenue (32%) increases” ⁽¹⁾ They also found that the UK has risen to become the second most important destination for investment in the coming year (the U.S. came top).

For AI to excel and flourish, the public and private sectors need to work together to deliver on the promise of AI and transform how we work and live. The strategy published by the government is a fantastic start and allows the private sector to have a clear view of where the government is headed.

We see the National Data Library as a hugely exciting initiative. If successful, that data has the potential to transform our lives, from healthcare all the way through to traffic jams and potholes!

And, of course, the regulation, safety and infrastructure required to power this transformation are vital cornerstones to giving us the computing capacity and confidence in what’s to come.

At Bytes, we have worked with the public and private sectors on various AI projects over the last few years. And while it continues to be early days, when the correct use case is found, the results are tangible and clear to see. But we have also seen many projects stall, often down to inflated expectations and poor data.

According to Gartner, at least 30% of Generative AI (GenAI) projects will be abandoned after the proof of concept stage ⁽²⁾ by the end of 2025 due to factors like poor data quality, insufficient risk controls, escalating costs, or unclear business value.

There are five key areas to consider to ensure your project succeeds.

1. Data quality

Using the right data

This will vary from organisation to organisation. Often, the temptation is to go for the “moonshot” use case. We recommend evaluating some potential use cases, reviewing their importance and ease and evaluating accordingly.

A relatively simple early win is a great way to show what’s possible and build a framework for future larger projects.

Data reliability

The foundation of any successful AI project lies in the quality of data. Reliable data ensures that AI models produce accurate and actionable insights. Identifying and curating high-quality data that reflects real-world scenarios and is free from biases and errors is imperative. Understanding your data and your data’s quality is one of the foundational cornerstones for success.

Synthetic data

Synthetic data can be a valuable resource in scenarios where real data is scarce or sensitive. Synthetic data is artificially generated and mimics the patterns and characteristics of real data without compromising privacy. This approach can be particularly useful for training AI models, enabling them to learn and make predictions without the risk of exposing confidential information.

2. Finding the right use case

Finding the right use case is not just about having the right data; it’s equally about understanding the people and the organisation. Each organisation has its unique structure, goals, and challenges. Therefore, it is crucial to align AI initiatives with organisational objectives and ensure that the chosen use cases resonate with the team’s capabilities and the organisation’s culture.

3. Technology

The recent hype for AI has largely been centred around GenAI; ignoring the hype and focusing on the right technology for your needs is essential. AI is a catchall for many different technologies, after all.

For example, machine learning focuses on training algorithms with data to make predictions or decisions without being explicitly programmed for the task, whereas GenAI specifically involves creating new content, such as text, images, or music, based on the learned patterns from the training data.

Organisations should evaluate technology options based on their specific needs and goals. Other factors to consider include scalability, ease of integration, and compatibility with existing systems.

4. Secure your data

Data security

Identification and protection of your data have always been important, but with AI’s ability to analyse enormous datasets quickly, it is now paramount that you know what data you have and that it’s adequately protected.

Organisations must comprehensively understand their data, including its sources, structure, and quality. This knowledge enables them to identify potential risks and ensure that AI models are trained on accurate and representative data. Understanding data lineage and provenance can also help maintain data integrity and traceability.

5. Adoption

Cultivating a culture of innovation

Successful AI adoption requires the engagement and support of stakeholders across the organisation. However, to ensure it’s successful and widely adopted, fostering a culture of trust and collaboration is vital, as there is a lot of excitement and fear about what AI can do.

We have seen that when experimentation is encouraged, innovation rewarded, and open dialogue encouraged, the feedback loop greatly accelerates adoption.

Training

Training and support to employees are crucial for overcoming resistance to change and ensuring the smooth adoption of AI technologies. Tailored training programmes can equip staff with the skills and knowledge needed to work effectively with AI systems. Ongoing support and resources can also help address any challenges that may arise during the transition.

AI revolution

AI can revolutionise every workplace, but like all technology, its deployment and adoption must be carefully considered. By focusing on data quality, identifying the right use cases, selecting appropriate technologies, securing data, and managing change, organisations can successfully harness the benefits of AI.

With careful planning and execution, AI can become a valuable tool for driving positive outcomes and transforming public services.