Reinsurers are investing heavily in AI for catastrophe modelling and pricing, but they are missing an opportunity to remove the administrative burden that drags down underwriters and operations staff, according to Mea Platform chief executive Martin Henley.
Speaking to Global Reinsurance at the Rendez-Vous de Septembre conference in Monte Carlo, Henley said the sector has built sophisticated analytics capabilities yet still relies on expensive talent to shuffle unstructured data such as slips, submissions, and bordereaux.
"The slips, the submissions, the bordereaux, all the unstructured data coming in. And then you've got expensive underwriting talent, or even highly skilled operations talent, just doing admin all day and moving things around," Henley said.
Data quality sits at the centre
Henley said the problem is not a lack of ambition among reinsurers. Instead, many have tried to retrofit generic technology into complex insurance workflows, which rarely works. The quality of data underpinning those processes is central to improving technology-driven value chains.
"If you don't get it right up front, everything downstream from there, the quality of the data gets worse," he said.
Softening market conditions have made this an operational priority rather than a nice-to-have. "As [the] market is softening, this efficiency point around understanding your data in lots of detail stops really being an optional thing," Henley added.
Where AI adoption is - and isn't - happening
Appetite for AI in reinsurance is growing, particularly for risk ingestion, data extraction, triage, and repetitive middle and back office tasks. But reinsurers remain cautious about letting AI influence underwriting decisions directly.
"Where we're seeing that people are still less comfortable, which is very understandable, is where you're letting the AI near decision-making, so near the actual underwriting," Henley said.
The nature of industry conversations has shifted significantly. Henley said discussions have moved from evangelising the technology to practical implementation: "It's not evangelising 'this technology could help you'. It's 'we know it can help us. How do we actually make it work in our environment?'"
This shift matters for professionals working across AI for Insurance, where the focus is increasingly on deployment rather than discovery. Henley also argued that insurance-specific technology is becoming more important, as generic AI struggles with complex specialty risks where explainability, auditability, and accuracy are essential.
Not a headcount play
Henley rejected the idea that AI adoption should be viewed primarily as a way to cut staff numbers. Instead, he framed it as a way to redirect expert employees toward higher-value work.
"Our view is this is an opportunity to get those expert people out of doing admin and help them to drive your business," he said.
That perspective aligns with broader trends in AI for Operations, where automation targets workflow friction rather than workforce reduction.
Why this matters for insurance professionals
For underwriters and operations staff in reinsurance, the message is direct: the administrative work eating into your day is increasingly seen as a solvable problem. The question is no longer whether AI can help with submissions, bordereaux, and data extraction - it's whether your organisation has the data quality and workflow design to make it work. Expect implementation discussions to accelerate as softening market conditions squeeze margins and make operational efficiency a competitive requirement.
Your membership also unlocks:
