Gavin Lillywhite: Breaking the headcount-growth equation

Xceedance is what’s known in the business world as a “high-growth company”.

Founded in 2013 by ex-Berkshire Hathaway India chief executive Arun Balakrishnan to “deliver insurance services and innovative technology solutions across the industry’s life cycle”, the firm now boasts over 5,500 employees across 350 global clients.

To date, Xceedance has fuelled its prodigious growth through headcount. Its operations now range from underwriting solutions to accounting, claims management to catastrophe modelling and property inspections to billing – and as its functions have scaled, so too have its staff.

Now, however, it wants to change how it does things.

“We’ve got a strong growth agenda and, in order for us to grow to the numbers we want to grow to, we can’t carry on doing it the way we’ve always done it,” explains Gavin Lillywhite, senior vice president for business development at Xceedance UK, Ireland and Europe.

“We can’t just add more full-time equivalents (FTE) and expect FTEs to keep doing more. Equally, clients are now expecting a different model as well.”

Lillywhite was speaking to Insurance Times at this month’s Managing General Agents’ Association’s (MGAA) annual conference, discussing how Xceedance was planning its evolution.

The first step, according to Lillywhite, is the build-out of a series of artificial intelligence (AI) agents aimed at improving business efficiencies and automating workflows. The suite of AI tools is known as Ian, short for “Xceed-Ians”.

He explains: “What we’ve been working on is augmenting the business model with an agentic platform to automate the process. We’ve built an agentic suite with a number of specific agents, that we can then implement in our own organisation.

“[Transitioning our operating model] means that we can help clients grow without a linear or exponential cost increase, in terms of servicing.”

Non-linear cost increases

Helping customers grow without a commensurate growth in Xceedance’s own headcount or the fees it charges is a topic that Lillywhite often returns to. It is a topic that reflects both the technological advances he hopes to capitalise on, as well as the current state of the market cycle.

“Let’s face it, generally the market is softening. By default, you end up with more submissions, and you have to triage more submissions just to be able sustain your numbers, let alone increase them,” he says.

“What you don’t want as a business is a linear cost increase [to triage submissions], bearing in mind you’ll probably never write half of that business.”

However, he’s keen to stress that the internal efficiency gains the firm hopes to achieve aren’t driven by the aim of reducing staff count and thus running costs, but rather with the goal of redeploying staff into more valuable roles.

“We’ve got 5,500 people. We’ll utilise our team members differently, because they’ll be able to manage more clients, because they move from a pure operational role to an oversight and supervisory role,” he adds, echoing a commonly held belief among AI-transformation professionals.

“The analogy I always give is it’s like going back 50 years to a car production line. People physically put everything together, bolted everything on, but now robots do the majority of work.

“You still have quality control, you still have engineers building it in the first place and you still have supervisors working along the production line – but the production job has changed.”

Public launch

After an internal stress test, the Ian agentic platform will be released publicly to insurance customers. The benefits of releasing a production-ready, internally-used tool are myriad, Lillywhite explains.

“Ian will be launched [to the public] shortly, but in the meantime, we’ve been building it into our own operations, which means that we can take out costs in our own operations,” he says.

“What we’re now saying to clients is we’ve built this for ourselves off the back of our insurance knowledge and operational experience, so we’re now looking to build agents for clients. Those clients can plug into what we already have, or if they have a specific workflow need, we can build something very quickly for them using the Ian platform.”

Moreover, the firm can use any internal successes and efficiency gains as a practice-what-you-preach endorsement of the product.

Lillywhite continues: “By building it internally first and testing it, we can then clearly articulate to clients that actually this does work, and this is what we’re doing already. That’s been an integral part to our growth strategy.”

So how will the Ian platform work and what can customers use it for?

Lillywhite explains: “The easiest way to think of it is like an app store. People ask us, is it complete? Well, is the app store complete? No – people add new apps to the app store every day and it’s the same principle.”

“We’ve built the foundational agents and now we’re looking to increase those agents as time goes by,” he concludes.

So now, with the technology firmly out of the cradle, the challenge for Xceedance is to prove that its AI agent platform can do more than make minor efficiencies gains – that it can break the firm’s long-held link between growing profit and growing headcount.