To say that the collections process is complex is nothing short of an understatement. Whether it’s navigating multiple systems, interpreting lengthy standard operating procedures (SOPs) or managing growing regulatory scrutiny, all are part and parcel of collections operations. Above all, there is the responsibility to handle high-pressure conversations and support customers through what is an incredibly sensitive and difficult period in their lives.
While arrears levels across both residential and buy to let have continued to stabilise and reduce, there’s no time for complacency – especially as the number of properties taken into possession continues to increase. In Q1, a total of 1,260 properties were taken into possession. While still below historic averages, that number only seems to creep up.
This is all against a backdrop of increasing volatility in the wider economy and on the world stage, with significant uncertainty around the long-term impact of the Iran conflict. Even if tensions do ease, the wider implications for inflation, interest rates and costs across the board are likely to be felt for some time yet.
As collections teams continue to operate in this increasingly complex and interconnected environment, it is only right to expect ongoing pressure across the market – particularly as lenders navigate the changing cost of credit and both household finances and mortgage affordability remain highly sensitive to changing sentiment and shifting interest rate expectations.
For teams balancing regulatory expectations, customer vulnerability considerations and operational efficiency, new technologies such as agentic AI are beginning to transform operations and reshape what good looks like in practice.
There is understandably a huge amount of discussion and activity around AI across financial services, but all too often, it is still theoretical or at best, entry level. In reality, AI is already delivering tangible benefits in high-risk operational environments like collections.
By leveraging cutting-edge agentic AI, teams are able to gather and interpret complex information, historic case data and procedural requirements in real time, providing agents with immediate guidance and recommended next steps.
Agents are able to do this without navigating multiple systems or manually interpreting detailed SOPs during sensitive customer conversations. Instead, agents are able to focus on supporting the customer and are guided to the most appropriate outcome.
The end result is real operational improvements, whether it’s reductions in handling times, less dependency on subject knowledge and far greater consistency across interactions and decision making. Just as important is increases in colleague satisfaction by reducing administrative burden and decision fatigue.
That is the critical point – it is all about enhancing the human touch, rather than replacing it.
Collections journeys will always require empathy, judgement and emotional intelligence. Customers in financial difficulty are often vulnerable, anxious and facing complex personal circumstances. No technology should remove the importance of human oversight from those conversations or that human-led interaction.
Instead, AI should support advisers by removing friction, delivering insights faster and helping ensure consistency and compliance in real time. This is particularly important within a highly regulated environment like financial services.
As exciting as the opportunities around agentic AI are, firms also need to approach implementation carefully and responsibly. Collections environments involve highly-sensitive conversations and even more sensitive data. There is strict regulatory obligations and significant conduct risk. Poorly governed AI deployment could create serious consequences for firms and customers alike.
This is why I believe many firms will increasingly look to trusted specialist partners to support AI adoption, rather than trying to build everything in-house.
Successfully deploying AI within collections is not simply about plugging in new technology or downloading a new app. It requires deep operational expertise, robust governance frameworks and an understanding of the regulatory environment collections teams operate within every day. This point cannot be understated – it is the combination of regulatory knowledge built through deep-domain expertise, combined with AI expertise and guardrails, built through exploration, experience and both internal and external implementation.
For providers already embedded in highly regulated servicing environments, AI presents an opportunity to combine decades of operational knowledge with advanced technology in a way that is safe, scalable and outcome-focused.
As economic pressures continue and customer expectations evolve, lenders will need collections operations that are more efficient, more consistent and more customer-centric than ever before. Agentic AI can absolutely help achieve that – so long as it is deployed correctly and backed by the right expertise.
In short, the firms that will succeed will not be the ones that adopt AI the fastest. They will be the ones that adopt it in the most responsible way.
