As government investment in new technologies grows, how do they decide where to begin?
For the UK Government’s Insolvency Service, which tackles financial wrongdoing, supports people in distress and works to maximise creditor returns, the challenge was not whether artificial intelligence could help, but how best to apply it.
The agency’s Digital Innovation Team believed AI could sharpen fraud detection, improve efficiency and enhance customer services. Yet the abundance of possible applications risked overwhelming any attempt at implementation.
Forging a plan
To chart that course, the agency turned to data and AI consultancy Aiimi.
Nathan Marsh, digital innovation lead at The Insolvency Service says that the consultancy gave us a strong “footing to start [its] AI journey.”
Beginning in late 2024, Aiimi embarked on a deep-dive analysis of how data was being used across the organisation. Over 20 hours of interviews were conducted, spanning seven directorates and 17 teams.
The research unearthed key themes defining the agency’s data use, highlighting both friction points and opportunities for transformation.
These findings formed the backbone of an AI use case assessment framework tailored to the agency’s specific needs.
The outcome was a three-year roadmap, narrowing down 80 potential AI applications to 29 promising uses, and ultimately prioritising five high-value solutions.
Among these are an external-facing chatbot to reduce workload and provide 24/7 customer access, enhanced fraud detection tools capable of scouring large volumes of financial data, and systems to help staff target and prioritise cases more efficiently.
According to Joshua Swords, head of data & AI engineering at Aiimi, the greatest challenge was filtering ideas quickly without losing sight of value.
“The primary challenge was managing a large number of ideas within tight time constraints,” he says. “It was crucial to quickly identify those with the greatest potential.”
Interviews with staff revealed the surprising breadth of how data is applied within the Insolvency Service – from public-facing policy consultations to highly confidential forensic investigations.
“This diversity highlighted both the versatility of the agency’s data and the wide variety of opportunities for AI to enhance operations,” Swords says.
A framework for the future
The AI framework that was developed has already proven to be reusable, says Swords. Aiimi has applied similar methodologies in sectors such as water and nuclear, and Swords sees broad potential across government and industry.
Joshua Swords, head of data & AI engineering, Aiimi
The framework evaluates each use case across four lenses: value, data, risk, and feasibility and cost. “This structured, adaptable approach ensures organisations can make confident, evidence-based decisions,” he explains.
Marsh agrees that the process has delivered more than just an AI roadmap. “Aiimi has provided us with a detailed roadmap that addresses our unique needs while having the flexibility to test and adapt as we go,” he says.
“Their research has helped us identify key areas for AI, but also given us broader insights into how we manage data, which is crucial to our wider work.”
Looking ahead
As the agency prepared to move from planning to delivery, both Marsh and Swords acknowledge that AI adoption is as much about people as technology.
“The potential gains are significant,” says Swords. “AI can automate repetitive processes, freeing staff for higher-value work and improving service delivery.”
But Swords stresses that to fully embrace AI, firms and government agencies need to open up to more training programs “to move ideas into implementation.”
“Change management is also crucial, as AI can feel intimidating and unfamiliar to staff, and integrating new tools requires careful guidance.”
For Marsh, “embracing AI aligns us with the government-wide drive for automation, while making it easier to do business with the agency and improving customer experience,” he says.
Aiimi will continue to support The Insolvency Service as it begins to build and implement its chosen AI applications. The hope is that, by collapsing an overwhelming array of options into a clear strategy, the agency is now on a path to long-lasting transformation.
