What practical AI adoption looks like for insurers

The potential of AI is different for every insurance business, but there are practical steps firms can take to unlock that value. Fiona Nicolson reports.

Despite widespread discussion around AI, many insurers are still trying to work out where the technology can deliver meaningful operational value. While some firms continue to explore large-scale transformation programmes, others are focusing on more immediate opportunities to improve underwriting, claims and internal workflows. 

In the latest Insurance Post webinar, AI in insurance: Moving from hype to practical impact, in conjunction with Hyland, Clare Knight, managing director of Bspoke Underwriting, and Jonathan Whitear, speciality sales for EMEA and APAC at Hyland, explored how insurers are approaching AI adoption in practice – from overcoming internal hesitation to identifying realistic use cases and preparing employees for change.

A point that came through strongly from the outset is that there is no blueprint for adoption. Each business is on its own individual ‘journey’ and must decide what form it will take. 

Hesitation and misunderstanding

Some insurers are still cautious about committing heavily to AI initiatives.

“Businesses are hesitant about AI because it is an overwhelming topic,” said Whitear. “It’s moving at such a pace that they are afraid to spend money on it at this stage in case they get it wrong.”

Knight said uncertainty and misunderstanding around AI persist across organisations, particularly where senior leadership teams are less familiar with the technology.

“Certain demographics understand AI better than others, but as with any mass change, it needs to come from the top down.

“It doesn’t matter if you have Generation Z employees able to use Claude code to build their own tools – they’re not the people in the business that are going to enact and empower change – and change is harder to implement, if you don’t have understanding at senior level, of where AI opportunities lie for the business.”

First-stage adoption

While AI developments are underway at Bspoke, which is beginning to roll out Microsoft Copilot, Knight believes that the main benefits will come later, arising from transformational changes and from embedding AI throughout the business: 

“Twiddling around the edges with AI isn’t where I see the big benefits. We can use Copilot for the minutes from meetings, which will save us a bit of time, but it’s not going to move the dial. The goal is true operational change, and that’s what I’m focused on.”

You can use what you already have and overlay agentic AI. It doesn’t have to be a huge transformation project.

Clare Knight, Bspoke Underwriting

Whitear said that learning AI basics provides a good launch pad, as long as there is a clear strategy: “It’s important to play around with AI, to develop the ability to use it beyond Copilot, but it should be part of a process, from where you can make big gains.

“Creating comfort with prompting is key, including understanding how it works and where it can be modified. Once the fundamental levels of comfort have been established, you can start to see what AI could do for your business.”

He added that insurers should view AI as a way of extending employees’ capabilities, rather than replacing them.

“AI is about shifting a knowledge worker’s information from their head into a prompt that then allows them to create work that they may not be able to do as effectively as a quick, scalable piece of IT can. That scares people, but they need to see this as an enhancement of their role.”

Whitear also said there is a key strategic question leaders will need to confront.

“The main point to consider from a board-room perspective, is: are we going to make this happen or is it going to happen to us?

“AI has the ability to change business models dramatically, so strategically, you want to be in control.”

Knight said Bspoke had decided early on to take a proactive approach. The company’s initial focus has been on improving understanding internally and helping employees become more comfortable using AI tools.

“The first step has been getting the team on board with AI and putting them through training, so that they understand the power of AI, and how it can help them.”

She also stressed that the technological upheaval is not as drastic as some might expect.

“You don’t have to get rid of your existing systems. You can use what you already have and overlay agentic AI. It doesn’t have to be a huge transformation project.”

Stumbling blocks

The discussion also highlighted the challenge of moving beyond early experimentation into projects that can demonstrate measurable business value.

“While a lot of people are comfortable with the early stages of getting to know AI, such as working with Copilot, progressing beyond that can be hard because businesses need to think about return on investment,” Whitear said.

“They have to consider where AI makes a difference in their business and where it makes fundamental improvements. That’s a board-level discussion.”

There’s a lot of ‘moonshotting’, when the real benefit of AI is at the other end of the spectrum.

John Whitear, Hyland

According to Whitear a shift in mindset is needed.

“AI is not just an IT issue: it’s more of a business transformation move than an IT development – and it represents a massive change to how businesses think about things.”

But over-ambition can also get in the way: “There’s a lot of ‘moonshotting’, when the real benefit of AI is at the other end of the spectrum.”

Keeping AI grounded in business needs

As AI evolves, firms need to keep their AI operations up to date and relevant. Knight says: “When I look to the future, I envisage having an agentic engineer embedded in my team, who understands the business and the systems that we’ve built, and will constantly be working with the team to move our AI operations forward.”

While acknowledging that the tech is advancing in leaps and bounds, Whitear cautioned that there is no pressure to continually enhance what you have: “It’s a bit like Excel. Most of us tend to use about 20 per cent of its capabilities.

“When a new version comes out, it doesn’t mean you have to use it or that it will improve your performance. The point is, not to ‘chase’ AI developments and instead focus on what the tech could do for you and how it could improve your business.”

Knight highlighted KYC onboarding as one area where AI could quickly reduce manual workloads.

“We provide capacity to UK cover holders for personal-lines business, and KYC is a massive part of the onboarding process. It involves reading a huge amount of documentation, which AI could do, if given the right prompts. This is a big time saving, taking the strain away from ‘the human’, and is one of the first things that we’ll be doing.”

The pair also considered where accountability will sit, where processes are fully agentic, with Whitear stressing that businesses are in the driving seat: “Companies choose what they’re going to do with AI and what they’re comfortable with. Its capabilities are not going to be imposed on them.

“How they build in checks in depends on their level of comfort, which may be based on whatever the rest of their industry is doing or what fits the culture of the company.”

The changing workforce

Alongside operational change, both speakers acknowledged growing concerns among employees about how AI could reshape jobs and career paths.

“I mentor people who want to enter the AI industry and recently helped someone with a school project. Their biggest concern was whether AI would take their job before they’d got it. That’s quite a worry for someone not in the workplace yet.

It’s an opportunity for us to be doing more interesting work and less boring admin.

Clare Knight, Bspoke Underwriting

“There are loads of courses you can do, but it’s a balance between understanding what the tech does and what the business process does. Being able to link those two things together is a tremendous skill. A prompt engineer, one of the new roles that we’re starting to see appear, would be really good at that.”

He pointed out: “When big tech changes happen, you have to change your skillset.”

Knight sees an opportunity to attract new talent to the insurance sector. “Being able to understand business process and know what prompt to put into the AI, is an up-and-coming niche in our industry.

“You don’t need to have 10 years’ experience in the insurance sector to do this. It may be beneficial if someone’s coming in with a new set of eyes and genuinely being curious about why things are done and what benefits they bring.”

While she does not expect hiring criteria to change significantly in the short term, Knight believes AI will gradually reshape the nature of insurance work.

“We’re still going to be recruiting for the same jobs with the same experience for some time. However, I think the quality of the roles on offer will be better and the jobs more attractive.

“With the introduction of AI, we’re still going to need people in the loop, and the nature of the job will move more towards human interaction. We’ll also find there are parts of our job we can now do, that we couldn’t previously, because we didn’t have the time or the bandwidth to see them.

“It’s an opportunity for us to be doing more interesting work and less boring admin,” she said.

---

View the Insurance Post webinar, AI in insurance: Moving from hype to practical impact, here.

Sponsored content