AI is becoming part of how insurance does business. It's already supporting underwriting decisions, triaging claims and flagging fraud risk, and that shift isn't going to reverse.
What matters more is what it means for the people doing the work. As AI takes on more of the routine, the value of holding deep technical expertise doesn't shrink, it grows, because someone still has to know when the model has got it right and when it hasn't.
AI is already part of the job
The efficiency gains from this are already measurable. Aviva has deployed AI models across claims that cut liability assessment times by 23 days and improved routing accuracy by 30%, alongside a meaningful drop in complaints, and results like that are why adoption across the industry continues to accelerate.
Why expertise matters more under AI, not less
Efficiency gains like this come at a cost if they're not managed carefully. As AI absorbs routine, entry-level tasks, PwC's research has found that skill requirements for roles exposed to AI are evolving around 66% faster than in other fields, while the day-to-day work that used to build technical judgement gradually disappears.
That leaves a gap that only deep, deliberately built technical knowledge can close. Complex claims, unusual risks and high-value policies still demand a level of judgement that comes from real expertise, not from a model.
The case for keeping a human in the loop
This is why a human in the loop matters, and why that human needs to hold real technical knowledge to do the job properly. Judgement only holds up if it's backed by real expertise behind it. As Rob Flynn, former UK Commercial Lines Chief Transformation Director at Intact, has put it, AI can improve efficiency, but human judgement is still what makes the difficult calls work.
Holding that level of expertise is what will position our members to play a defining role in the workforce AI is shaping, qualified to check AI's outputs, catch what it gets wrong, and take responsibility for the calls that still need a person to make them.
The real challenge ahead
The challenge for the industry now is balancing technological transformation with the human experience that underpins trust in insurance. Getting that balance right depends on the same thing it always has: professionals who hold the technical knowledge to make good judgement calls, whether or not AI is involved in reaching them.
Building that knowledge
Our new AI in Insurance programme is built around that reality. It's designed specifically for how insurance actually works, not as a general AI course with industry examples added on top. Over six weeks, participants work through real insurance case studies covering where AI is already being used, what can go wrong, and what governance and regulation, including the EU AI Act, expects of them.
Participants build the practical skills that separate professionals who can direct these tools from those who simply defer to them, including how to structure prompts properly and how to interpret AI outputs critically rather than accepting them at face value, and rather than a formal exam, they leave with things they've actually built and applied to their own role.
Full details, including dates and fees, are available here.