All insights
Visual representation of an AI native future

Inside the machine: How Willis Re is building an AI-native future for reinsurance

Matthew Whyte, Managing Director, Global Specialties, Willis Re

Willis Re is building the market’s most advanced reinsurance broking platform that puts clients first, says the firm’s managing director, reinsurance broker Matthew Whyte.

In a very short space of time, Willis Re has reached technical equivalence with its peers in analytics capabilities. We are now focused on differentiating through enhanced client services. As agentic AI adoption accelerates, we can innovate without the constraints of legacy systems. While much of the market is focused on using generative AI to extract value from existing data repositories, we are building a truly AI-native platform. This platform will employ AI agents where appropriate, complemented by human expertise to govern, validate and continuously improve decision-making.

The how

Determining the strategy is straightforward but the engagement of capable and experienced partners is essential to implementation.

An important part of the vision is that all data can be stored and retrieved from one interactive ecosystem where data integrity is paramount. We have engaged Snowflake to deliver our cloud-native data warehousing platform. It’s important to recognise that the term “data” now includes the written word due to the development of large language models (LLM).

For the buildout of our AI-native platform, we have engaged with sister company Newfront, which is providing its design and engineering capabilities to add to our reinsurance broking expertise to build something totally new – a client-centric native AI broker platform. No more sales tools manifesting as “client relationship management” (CRM). No more unintegrated management reporting tools. No more clunky email filing systems. No more disconnected fiduciary systems. All the information integrated into one platform with instant and live recall.

So what?

“It’s all very clever but what does it do for us?” We have listened to our clients and markets to hear what they want. This is not an “efficiency” play, though efficiencies are there. The focus is on providing more insight and enhanced service to clients while making it easier for our markets to deal with us.

Our reinsuring partners will not require access to yet another broker-operated electronic placement platform. Communicate with us in the way that suits you. AI will automatically recognise the quote or offer for what it is. No more complex login processes or filling out electronic forms.

Our clients will be better served with:

  • a significantly swifter turnaround of analytic output using data ingestion techniques that will cleanse data, develop submissions and automate stochastic and deterministic modelling;
  • vastly improved market insights using contextual data from all market correspondence;
  • instantly available data points for peer analysis;
  • the capability to evaluate reinsurance structures against whole platform capital models; and
  • the end-state of secure client and market portals that enable a live view of our trading relationship. Premium and claim statuses, marketing and endorsement progress, balances, submissions. A complete view of all interaction.

So when?

The short answer is: by the end of the second quarter of 2027 for complete functionality but much of that described above will be online sooner.

Unlike traditional IT engineering where vast sums are committed to getting it right the first time to avoid rebuilding and the consequent lengthy time delays, AI-native builds are iterative by nature. Long planning stages are eliminated as AI instruction can be refined and changes made instantly with no cost or time implications. As such, development is always ongoing as we imagine new capabilities and functionalities to support our clients and markets.