The impact of AI on work has been a hot topic for years, and became even more heated since the World Economic Forum started predicting that generative AI would displace an unprecedented number of jobs. But on the ground, what happens when companies start onboarding AI?

On the sidelines of the AWS Summit Singapore 2025 last month, People Matters asked insurtech startup bolttech, which is rolling out a generative and agentic AI upgrade to its call centre platform, about how their own use case is turning out. 

bolttech's Asia CEO, Philip Weiner, said that the objective of bringing in AI in the first place was to get the company in a position to scale. It is already considered a high-growth company: founded in 2020, it was valued at more than US$1.6 billion (S$2.15 billion) by 2024.

The use case is a human one

Customer experience emerged as one of the more significant use cases for bolttech, as the company provides insurance for mobile devices that includes support via voice call or chat. It is also the key metric by which they are measuring the success of the AI upgrade, Weiner said.

"Customer experience is something that we measure very, very closely on a daily basis. Whether it's a generative AI helping the customer, or a human doing the interacting, customer experience is paramount."

He also highlighted technical metrics that feed into customer experience, such as performance on latency - not having too long a silence before delivering a response to a customer. And of course, cost effectiveness and scalability of the technology have to be at acceptable levels.

So far, Weiner added, the results have been positive. The bot responds faster and more accurately than human customer service agents, and when it has to hand over to a human, it does so in a way that transfers knowledge about the case to the person.

"We're not drastically different from other contact centre operators in things like average handling, time, speed to answer. We track all of those metrics. But when a human's interacting with a robot, do we care as much about average handling time? What we actually want is the problem solved. We want the customer query or complaint dealt with in their first interaction," added David Lynch, bolttech's group chief technology officer.

What happens to the human agents?

bolttech isn't planning to replace all its human customer service agents just because it has AI on board. But the kind of work those human agents do is becoming very different, Weiner said.

"What the team of humans is working on becomes very different. They are able to, first off, cover a much wider range of information than they used to, and they're dealing with much more complex, intricate, individualised situations that actually require them not to be someone who's the best at following a process."

The process, he explained, will have already been worked through by the AI before the query is escalated to the human agent; meaning that what the human agent receives, and needs to be able to navigate, is going to be a gray area.

"Understanding - empathy with the customer - comes up a lot in this space. Something has happened that was unexpected, and the standard process says for example, I should hand you a bottle of water, but I have the empathy to know that's not going to work in this case, and then I have to use my judgment. And so it is really transforming the skill: from process follower to judgment and navigating uncertain areas, and that is changing how we are looking at upskilling and training the contact centre staff."

Lynch added a little titbit to the conversation, perhaps tongue in cheek: considering the importance of customer experience in the business model and the need for a higher tier of these softer, more adaptive skills, bolttech just might be looking at hiring a Chief EQ Officer or similar role at some point in the future. It seems an appropriate counterpart to the technological implementation: an EQ upgrade to balance out the AI solution.