Following a session at the Money Advice Liaison Group (MALG) Conference 2025 entitled ‘Artificial Intelligence – Technology in Tandem with Humans’, we identified a gap in knowledge on the impact of generative AI on self-serve debt advice now and in the longer term.
MALG, in partnership with Wyser (whose founder and CEO Mark Pearce was on the panel at the conference), has secured funding from Money & Pensions Service (MaPS) to undertake a research project entitled: ‘Exploring Safe & Effective Self-Serve in Regulated Debt Advice in the Generative AI Era‘. The research is being conducted during January-March 2026 and the outputs from it will be launched at a virtual event on Tuesday 31 March 2026, 10-11am – register here.
The Research: Taking Part
The research is being carried out by Wyser (whose mission is to develop AI products that make the world a better place) from January to the end of March 2026. Debt advisers and their clients; adviser managers or supervisors; and colleagues involved in advice quality, assurance or oversight have been taking part in one-off 60-minute virtual interviews and we have now launched a sector-wide survey.
The survey takes just 15 minutes to complete and is open to anyone working across the debt landscape. It is open until 5pm on Friday 13 March 2026 – please click here to complete the survey now!
A consumer survey will also be launched soon.
Research Brief: Executive Summary
Generative AI tools, including chatbots, are increasingly accessible to people seeking information and guidance about debt management. These tools present potential opportunities to improve access, convenience and early engagement with advice. At the same time, their use within regulated debt advice raises significant questions about safety, trust, suitability and the risk of harm.
This research explores the opportunities, limitations and risks of AI-enabled self-serve tools in the context of regulated debt advice. It seeks to identify the conditions under which such tools may be used safely and effectively, as well as circumstances in which they must hand off to human advisers.
The work is exploratory and foundational. It does not aim to evaluate or certify a specific AI tool and it is not intended to replace regulated human advice. Instead, it builds an evidence base that reflects how consumers behave when interacting with AI tools, how advisers and managers assess risk and regulation in practice and where alignment or tension exists between consumer expectations of AI-enabled self-serve tools and professional real world knowledge of working within the sector.
Findings will inform a set of practical, sector-facing outputs including an evidence report, a safe-use framework, reusable templates and open license research tools. The research is delivered by Money Advice Liaison Group (MALG), in conjunction with Wyser, and is intended to support learning and responsible experimentation across the debt advice sector.
The Question
‘What are the opportunities, limitations and risks of self-serve in regulated debt advice in the era of Generative AI, and what conditions support safe and effective use?’
Why this question?
- The Sector can’t afford to ignore it: Clients are already turning to unregulated tools like ChatGPT for debt help. If we don’t ask this question now, the sector risks being left behind with unsafe, inconsistent advice filling the gap.
- The potential benefits are wide ranging: If self-serve can be made safe and effective, it could free up advisers for the most complex cases, reduce waiting times and give millions faster access to help. The prize is a transformed, more resilient debt advice sector.
- Building the foundation for future policy and funding: This question creates the evidence MaPS, FCA, and providers need to make the big calls on digital debt advice. Without asking it, future strategies will be built on guesswork rather than hard evidence.
Deliverables
- Evidence Report: A concise synthesis of findings from a rapid desk research review, client task testing, adviser focus groups and a sector survey.
- Safe-Use Framework for Generative AI in Debt Advice: A practical tool defining when AI-enabled self-serve is appropriate, for whom, and with what safeguards.
- Templates Pack: A set of reusable design assets that providers can adopt immediately.
- Dissemination Package: A co-branded presentation deck & summary briefing.
During the research, it has also emerged that the following additional deliverables would be of use:
- Adviser Capability & AI Literacy Guide: A practical guide designed to equip advisers to work safely alongside AI.
- Turning Organisational Data into Safe Self-Serve AI Guide: A practical guide to help organisations use internal data safely for AI-enabled self-serve.
