One Core, Many Countries: Why Harmonisation is the AI Multiplier.

The fastest way to scale AI across a regional or global insurance group is not a better model. It is not a bigger AI team. It is a core insurance platform that is essentially the same in every country. 

Core modernisation is an unfashionable topic, slow, expensive, and often political. AI is the fashionable one. But core systems and AI are far more connected than most group executives realise (or admit?). 

Here’s the problem with running fragmented cores across a regional insurance group. An AI agent built in Country A — for example, an FWA assessor for medical claims — has been integrated into the local policy admin and claims system. It has been trained and evaluated on Country A's data.  

Now you want to roll it out in Country B. Different policy and claims data model. Different workflows, different product architecture, different integration patterns, different master data definitions. What was supposed to be about reusing quickly became rebuilding. 

Repeat this across six countries, and AI rollout in a regional group becomes a one-country-at-a-time grind. The compound effect that makes AI powerful across large data sets is exactly what gets killed. 

On a harmonised, multi-country core platform, this changes. The integrations are standardised. The data model is the same. The workflows are largely harmonised, with local variance handled through configuration rather than custom code. AI agents can be designed and trained on the group's full data footprint rather than one country's (where it makes sense). Governance, compliance and security can be centralized to become tighter. 

It is not just about faster and better outcomes. Importantly, it is about cost efficiency as well. Almost all insurers face significant IT investment backlogs. Budgets are already stretched and often struggle to keep up with inflation. Required investments in AI fall short, as the budget (and change bandwidth) are often not there. 

We have been working with insurers on harmonised, multi-country operating models. We see efficiency gains in hosting and maintenance of up to 70%. We have seen implementation time and effort reductions by 40–50% for subsequent countries. These figures are before AI agents enter the equation. Once they do, the same harmonisation compounds: every agent built on top inherits the reuse. 

The implication for regional and global insurers: Core modernisation matters. But it should not just be modernisation, but rather be combined with harmonisation. The group that runs on one flexible, harmonised core will out-compound the group that runs on a dozen local cores. 

If you're a group CTO or regional transformation lead, and your AI strategy and your core strategy are still being discussed in different meetings, that is a governance issue to fix.

-Bill Song, Peak3 Co-Founder and CEO