A major new study released by Money Advice Liaison Group (MALG) and AI innovation firm Wyser provides the clearest picture yet of how generative AI could reshape the UK’s regulated debt advice landscape – but only if deployed with precision, humility and human oversight.
Funded by the Money & Pensions Service (MaPS), ‘Exploring Safe & Effective Self‑Serve in Regulated Debt Advice in the Generative AI Era’ lifts the lid on how AI‑enabled tools might support earlier engagement, improve access and reduce adviser burden. But it also highlights deep structural complexities, widespread vulnerability and a public trust gap that mean self‑serve AI cannot simply be ‘switched on’ without robust safeguards.
The result of the research carried out with sector professionals and consumers from January to March 2026 is a sector‑first, sector-led evidence base; a Safe‑Use Framework for Gen AI in Debt Advice; and a suite of practical tools for organisations exploring AI adoption.
The Safe-Use Framework encompasses a three-level Debt Advice AI Roadmap, giving the sector a shared vocabulary for both current applications and future developments. This could also be used by regulators to develop proportionate guidance that supports innovation while maintaining consumer protection.
Mark Pearce, CEO and Founder of Wyser, comments:
“What this research makes clear is that the debt advice sector’s expertise must drive how Gen AI tools are built, not the other way round. AI can help, but only if we focus first on where it adds genuine value: cutting the admin that consumes adviser time and extending reach through carefully designed self-serve tools for the right clients at the right moments.
“It’s been a pleasure to collaborate with MALG and the wider debt advice sector on this project. The resources we’ve created aim to help organisations navigate the fast‑moving AI landscape without compromising client safety or advice quality. We see this as the starting point, not the finish line, and we are keen to support organisations in taking the next steps from evidence into implementation.”
Peg Alexander, CEO, Money Advice Liaison Group (MALG), says:
“We’re proud to have led on this ground-breaking piece of research, which is hopefully just the beginning of a longer-term project to ensure that the debt advice sector can keep pace with the rapidly evolving world of Generative AI while maintaining the principles on which it is built.
“Many organisations are operating with sustained pressure from rising demand and complexity but the sector’s structural challenges, the prevalence of deficit budgets, decreasing funding of free advice services, the administrative burden of regulatory compliance and the attrition of experienced advisers, cannot be solved by technology alone. We’re committed to leading on a collaborative way forward, working with stakeholders across the whole debt landscape.”
Anna Hall, Corporate Director for Debt Advice, Money and Pensions Service, adds:
“Through the Debt Advice Transformation Fund, we’re looking to support the sector to adapt, improve and respond to emerging challenges by testing ideas, building evidence and developing shared understanding.
“What made MALG’s proposal stand out was that it didn’t start with a solution; it started with practice reality, with the aim of helping the sector move forward thoughtfully and not just reactively. It has given us a shared evidence-based understanding of what the safe and appropriate use of AI actually looks like in a regulated debt advice environment, laying the foundations for future development across this sector.”
Key Findings:
- Debt advice is too complex for simple automation
The study confirms what advisers have long known: regulated debt advice relies on layered assessment, professional judgement and relational skills honed over years. Gen AI cannot yet replicate the nuanced decision‑making required across benefits, housing, enforcement and safeguarding.
- Vulnerability is everywhere – and often invisible at first contact
Advisers stressed that hidden vulnerability typically emerges only after trust is built. While AI may support low‑complexity tasks, it must not undermine the development of human expertise or miss critical red flags.
- Most clients arrive in crisis
From bailiff action to court letters, people often seek help at breaking point. Participants questioned whether AI tools could safely de‑escalate crises, especially when a confidently wrong answer could cause harm.
- Debt rarely exists alone
Housing, benefits and debt issues are deeply intertwined. Self‑serve tools that focus narrowly on debt risk missing wider stabilisation needs, particularly given high drop‑off rates between advice types.
- The financial statement remains the ‘beating heart’ of advice
Budget construction requires challenge, interpretation and future‑risk assessment. While AI can gather initial information, advisers were clear: strategic budgeting still demands human judgement.
- Advice is a relationship, not a transaction
Debt advice often spans months or years. Follow‑up with creditors, reassurance and re‑evaluation are core to the role – and difficult to replicate in a self‑serve environment.
- Client use of Gen AI is lower than expected – but rising
Despite media hype, advisers reported limited client use of AI tools so far. But emerging cases, including AI‑prompted bankruptcy decisions, show the trend is shifting.
- Public trust in AI for financial issues is low
The YouGov survey revealed that 72% of the public lacks confidence in AI financial advice, and over 80% prefer a human adviser. Regulation and human review were the strongest trust‑builders.
- The sector is divided – but not resistant
Larger organisations feel urgency; smaller ones are cautious. Across the board, contributors agreed that debt expertise must shape AI design, and that adviser‑assist tools offer the most immediate value.
- Adviser‑assist AI is the strongest early win
Tools that reduce admin (transcription, case notes, credit report analysis, legislation decoding) received overwhelming support, with 82% of survey respondents rating them suitable now.
- Self‑serve suitability varies dramatically by organisation
Client bases differ widely. For some organisations, self‑serve may never be appropriate. For others, it could support digitally confident clients effectively. There is no universal model.
- Accuracy and jurisdictional sensitivity are non‑negotiable
Debt advice changes constantly, and Scotland’s framework is entirely separate. Any AI tool must have robust update mechanisms and clear routing to avoid outdated or incorrect guidance.
-ENDS-
Find out more and download the evidence report and resources at: https://malg.org.uk/genai_debtadvice_research/