Small team, huge value: how an AI agent put internal audit at the forefront of technological change at The Exeter 

“If you’re too busy to think about AI, you’re the one who needs it most.” That is the message from Daniel Wilde and Samantha Preen, who developed an AI agent to benefit a small internal audit team using CoPilot. Their work won the Chartered IIA Internal Audit Innovation award in 2025 – and provides lessons for every busy team with no specialist AI knowledge. 

The Exeter is a health and life insurance mutual society with a small internal audit team of just three people, including an apprentice. Wilde, who was Head of Internal Audit at the time, described himself as having “lots of ideas, but little idea how to implement them”. 

He admitted it would have been easy to assume they did not have the resources to implement an AI system. This did not stop him and Preen, Senior Internal Auditor, from setting out to overcome their limitations, build their own AI agent, and revolutionise their processes. 

Even more impressively, they did it on a budget of £25 (plus a lot of hard graft and rapid learning). 

They began by joining an AI Working Group that The Exeter had set up to explore possibilities across the organisation. This put internal audit at the forefront of the organisation’s AI developments. 

Wilde then looked to see what he could learn from contacts in previous roles, and he and Preen searched for AI groups and advice sessions hosted by consultancies. They submitted a business case for a CoPilot Pro licence to trial their ideas and set to work. 

“We saw that AI developments were already happening and that if we were not part of this change, we would become less relevant. You can’t go into a management meeting and have nothing to say about AI and how the business is using it,” Wilde explains.  “By getting ahead, we increased our function’s reputation internally and the organisation’s reputation externally.” 

Preen agrees. “It’s about being a strategic adviser,” she says. “We need to leverage resources as best we can, so part of presenting our audit findings should include suggesting how we can use AI further to gain more insights.”

Building I.A.I.N 

She set about learning how to build an agent that could help them gather information to assist audit planning. She learned about prompts and how to constantly refine these to gain accurate, reliable, and useful suggestions – while still completing her scheduled audit work. 

Online training videos proved helpful, and colleagues and peers shared their experiences. However, much of the work involved trial and error and imagination. When they encountered hitches – for example, when the system failed to pick up their internal controls correctly or couldn’t access PDF content – they redrafted their prompts, reformatted their data, and carried on. 

They communicated their progress and what they learned to the AI Working Group and created a video demonstration for the board and Exco. They called their new agent I.A.I.N. to personalise it and support their communication. 

 “We had to make it clear that AI is not just about doing our job as it is now, it’s about doing a better internal audit,” Wilde says. 

Despite the extra work, they completed their internal audit plan on time and identified potential time savings they could gain by using I.A.I.N. in the future. In 2025, they planned to save 80 days on their 2024 plan, which they invested in additional follow-up audits, agile status reviews, and a dedicated data analytics audit.   

I.A.I.N. made processes faster and more efficient by ensuring that all relevant data was included in audit plans, summarising findings from working papers, and enabling the team to focus on more audit work. 

It also put the internal audit team at the forefront of AI development in the organisation. Wilde says it helped the team to build stronger relationships with stakeholders and inspired other functions, such as compliance, to experiment with AI tools. 

More widely, Wilde and Preen believe their initiative should encourage other smaller internal audit teams to build their own AI tools. They argue that any team can use existing tools, cheaply and effectively, to enhance their audit processes and add real value to what they do. 

“It’s about enhancing the way we do our job,” Preen says. “It’s a tool, just as a computer is a tool. You need to feed in ideas and direct it, and control the output, and you keep ownership of the process. But if you do it well, it may give you a different perspective and help you to think differently.” 

AI can be a useful sounding board in smaller internal audit teams that lack experienced colleagues to challenge and debate ideas. The CAE role can be lonely, and AI could ensure leaders consider all angles and stay on top of large quantities of data. 

“If you ask AI to do something that you are also doing and compare your findings, it can throw up areas you have missed,” Preen suggests. 

It’s essential for all audit teams to move on and change the way they work, Wilde and Preen add. “We’re seeing a shift in what boards and regulators want from internal audit, and using AI helps you provide the advice, insight, and service they require,” Preen says. 

Challenge is important, adds Wilde. “We need to challenge the business, and we need new people to come into the internal audit profession with ideas. If I hire a new person, I want them to challenge me and ask why we do things in a certain way. If we don’t challenge and drive change across internal audit and the business, who will?” 

Future potential 

Preen and Wilde are constantly looking at where I.A.I.N could help them in the future. Now in a new role at another company, Wilde is thinking about documentation and how AI could compare templates and check when policies have been updated and how regularly this is done. 

“You could use AI to identify who owns each policy, whether it has errors, and where it may need updating,” he suggests. “It means you start an audit in the best place and move beyond basic testing. It creates a benchmark so you can focus on the important stuff, such as the content of policies and whether they are pitched at the right level.” 

Preen adds that AI can continuously monitor areas that previously relied on sampling, such as payroll, and highlight anomalies as they occur. She wants to reach a point where all the relevant data from continuous monitoring of key controls is available as soon as she starts an audit.  

This shouldn’t just involve audits you’ve done in the past, Wilde adds. “AI enables you to look at things you’ve never thought of doing before.” 

“For example, if you are in insurance, you could analyse all the people who start an application process on their phone, but run out of battery – what does this tell you about this cohort? Are they risk-takers because they start with a low battery level? Do they come back if they crash?” 

Given an open question, AI may come up with wild suggestions that seem mad, or revolutionary, Wilde explains. “If you own a paperclip business and just ask AI what you should do next, you don’t know what its logic might lead to – it could have a limitless butterfly effect that transforms your business, or causes chaos.” 

Protect from the start 

New tools require new protections, which must be in place from the start. Culture and governance are vital elements, Preen warns. But this is an opportunity for smaller, cohesive teams. 

“If you have a good culture, strong data governance, and high-quality data, AI gives small internal audit teams a much greater opportunity to be strategic advisers and work at the highest levels,” she says. “They can do the same level of work as large teams in a way that has been more difficult in the past.” 

Ethics will become ever more important – and internal auditors are in a strong position to lead ethics discussions. “We need people who can audit culture, understand ethics, fact-check, challenge misconceptions, and cut through misinformation,” Preen argues.  

She has also used AI to analyse how the language in a report aligns with her objectives and with the culture and strategy of the business. “It suggested some alternatives that were a better fit, but the ultimate choice was mine,” she says. “I know I’m better at ideas and collaboration than at writing reports, so it was really useful.” 

Working with AI to improve the impact and subtlety of your communication could make you more aware of when others use AI to influence you, she adds. Like all powerful tools, AI can be used for good and bad purposes. 

Wilde agrees. “We saw in Covid that if you look at the data in the wrong way, you make the wrong decisions. AI enables access to data and helps you to consider all the factors to make better decisions. It helps you look at details in a way that may lead to completely new solutions.” 

This is just the start of the journey, he adds. AI will increasingly do more than simply replace what we already do faster and more efficiently, and this is why internal auditors must understand it and develop with it.  

The Award 

Winning the Chartered IIA Award raised the profile of what Wilde and Preen were doing, and they were asked to give a presentation at the Internal Audit Conference in October. They have also talked to people on the Chartered IIA’s Data Analytics Working Group.

“It was great to get so much positive attention from peers and those in much bigger internal audit functions,” Wilde says. 

“It’s cool bringing internal audit to the forefront of exploration and corporate discussions,” he adds. “For too long, internal auditors were seen as the people with the clipboards. This is about spotting opportunities and driving change. It’s exciting.” You can hear more inspirational stories like this one on the night.