Chris Hague: AI, Agents and the Human in the Loop

In this episode of AI Unplugged, Martin talks with Chris Hague, Managing Director of Bill Gosling Outsourcing UK, about how the collections agent’s job has changed since they both started on the phones. Chris spent around a decade in banking and a decade in global debt purchase before moving into UK debt collection. He explains why the current AI cycle will make earlier shifts like power dialers and digital self-service look small, how AI in training and on the floor, from agent assist to analysis of every call, has shortened the time it takes new agents to become competent, and why the result nobody planned for was lower attrition. The conversation then turns to a harder question: if AI absorbs the entry-level work, how does the industry develop the experienced people it needs for its most complex and vulnerable customer conversations?

What You’ll Learn

• From the phones to Managing Director: How Chris went from collections calls for an American bank at 27 to running Bill Gosling Outsourcing’s UK business, via team leadership, credit risk and ten years in global debt purchase.
• Why this AI cycle is different: How it compares with power dialers and the move to web portals and self-service, and why Chris expects it to dwarf both.
• What AI looks like on the floor today: Agent assist, early automation of repetitive tasks, and AI analysis of 100% of calls that feeds real-time coaching back to agents.
• The benefit nobody put in the business case: Why better-supported agents stayed longer, and why attrition is now one of the key measures Bill Gosling tracks.
• Why the agent’s job is harder than ever: Regulatory controls, Consumer Duty and the scale of customer vulnerability demand a different skill set from the days when agents were measured on cash collected.
• The junior pipeline dilemma: If AI takes over entry-level work, where do tomorrow’s experienced agents come from, and how segmented roles, longer training and AI roleplay could help.
• Lessons from automation rollbacks: What Morrisons putting staff back on its tills and Klarna’s U-turn after replacing staff with AI say about deploying technology without a clear problem to solve.
• How clients are approaching AI: Weighing concerns over data, customer engagement and reputational risk against the fear of missing out, and what that could mean for what they outsource.

Key Takeaways

• Support drives retention. AI in training and real-time support on calls shortened time to competency and, unplanned, reduced attrition. That knock-on value was never part of the business case.
• Agents see support, not a threat. Chris has not seen fear on the floor, because the AI agents work with today helps them do the job.
• The human in the loop is getting more skilled. AI takes the repetitive, mundane conversations; people handle bereavement, vulnerability and complex cases.
• Life experience matters. The hardest conversations call for experience you cannot expect from a 20-year-old, and the role is not right for everyone.
• Plan for the backfill problem now. Replacing entry-level roles with AI breaks the path that builds senior people. Segment roles, extend training and use AI for roleplay.
• AI must solve a defined problem. If it does not fix a problem, improve a process or improve an outcome, it is just pulling a trigger, and customers may push back, as Morrisons found.
• Clients move at different speeds. Some test early; others wait for peers to make the mistakes first. Culture and regulation, UK versus US, shape the pace.
• Listen before you sell. Deep domain expertise, a range of options matched to each client’s risk appetite, and “two ears and one mouth” define the BPO’s role.