A year ago, most conversations about AI in insurance started with curiosity. Today, they often start with anxiety and a lingering sense of not quite keeping up. Every week seems to bring another announcement, another capability, another prediction about how quickly the technology is advancing. Even those of us who spend our days immersed in the insurance market can struggle to keep pace. The gap between what seemed possible 12 months ago and what is possible today is remarkable.
That creates excitement, but also a degree of unease. The MGA sector has never been afraid of technology. In many respects, entrepreneurial businesses have always been quicker to embrace new tools than larger organisations burdened by legacy systems and processes. It is one of the reasons the sector has been able to innovate so effectively across specialist and emerging risks.
What feels different this time is the scale of the conversation. Previous waves of technology largely changed how we worked. AI has the potential to influence how we think, how we make decisions and how expertise itself is developed. That deserves a little more reflection than some of the current debate allows.
The opportunities are obvious. Underwriters spend too much time navigating administrative tasks, reviewing information that could be analysed more efficiently, or performing tasks that add little value to the customer-broker relationship. Few would argue against using technology to remove that friction.
Across the MGA community, there is genuine excitement about what AI could do to improve productivity, enhance data analysis, support fraud detection, and accelerate decision-making. The supplier community is investing heavily in solutions designed to make underwriting, claims and operational processes more effective. Much of that innovation is welcome.
Yet I find myself returning to a different question. What happens to expertise when parts of the learning journey disappear?
Every experienced underwriter can point to moments early in their career that shaped their judgment. Hours spent reviewing submissions. Discussions with colleagues. Mistakes that became lessons. Exposure to unusual risks. Gradually, experience accumulates and instinct develops.
If technology removes some of those stepping stones, how do we ensure the next generation acquires the same depth of understanding?
That is not an argument against AI. Far from it. It is an argument for investing just as seriously in mentorship, coaching and talent development as we do in technology itself. Insurance remains a people business. The product may be a policy, but what customers, brokers and capital providers ultimately value is expertise. The ability to make sound decisions in uncertain circumstances has always been the industry's defining skill.
There are other questions too. Security, governance and accountability all become more important as adoption accelerates. Most organisations are still establishing where the boundaries should sit. How much reliance is appropriate? Who is accountable when decisions are influenced by AI-generated outputs? How do we protect sensitive information in an increasingly connected environment?
These are not reasons to slow innovation. There are reasons to approach it thoughtfully. What encourages me is that the MGA market has a history of balancing innovation with pragmatism. It is a sector built on specialist expertise, entrepreneurial thinking and close relationships. Those qualities will remain valuable regardless of how sophisticated technology becomes.
Perhaps that is the real challenge ahead. Not whether artificial intelligence can transform insurance. It clearly can. The more interesting question is whether we can embrace that transformation while preserving the judgment, relationships and expertise that made the industry successful in the first place.
I am looking forward to exploring that debate further at the MGAA Annual Conference on 7 July, where, rather than discuss the technological AI advancements, we want to debate the consequences on the industry; how we can bring our people with us, what the recruitment landscape looks like, and whether there is enough focus on staff retraining. This conversation, I suspect, has not been sufficiently aired, which is precisely why it deserves our full attention.
