Generative AI is poised to revolutionize industries, and insurance is no exception. The journey, however, is fraught with challenges. To explore these dynamics, I recently interviewed Paolo Cuomo, Executive Director at Gallagher Re, author of “Quantum Computing: Why You Should Care,” and an insurance industry veteran. He offered unique insights into how AI can transform insurance—from enabling more informed decisions to elevating efficiencies for professionals at all levels. He also highlighted the role of leadership, foundational data work, and change management in building this new AI reality.
In what areas can the insurance industry benefit the most from adopting AI?
AI is one of the most transformational technologies we've seen. It will impact everything, and in the short to medium term, it will benefit efficiency by helping insurance experts spend more time on deploying their expertise because it removes tasks that don’t need their expertise.
Take the idea of the Digital Sherpa: currently, humans still make better decisions than machines, but there are multiple ways the machines can help enhance that human-centric decision-making. In the near to medium term, the digital Sherpa helps experts by providing access to all available information on a topic. AI can analyze every placement and claim in your company, along with public data, to inform decisions. It might say, "Here's a similar claim from our Australia office," or "Here was the reason a similar risk was declined five years ago." This way, experts combine their experience with comprehensive data for more accurate decisions.
What needs to be in place in terms of data, talent, and leadership for insurance companies effectively take advantage of Generative AI?
First, both the technology and the data have to be in place—and the data must be clean and usable, which is essential for training the machines.
When it comes to talent, tech and data professionals are in demand in every company today. Insurers and brokers are competing for that talent not just against other insurance companies, but also against retailers, telco firms, shipping firms, internet firms—every industry globally seeks this scarce pool.
Finally, there is a huge change management exercise involved. Consider the actual underwriting, broking, or claims side of things as the principal frontline areas of the business. Now, think about the efficient impact on people when simply upgraded to the latest version of Windows or switching underwriting systems—they struggle to find the right buttons or documents for days or weeks. Many firms will have already introduced an internal AI tool but find that only 10% of the business is using it, and only twice a week. We are bad at engaging with new technologies because they often do not provide enough improvement quickly enough to justify the effort to learn how to engage with them.
What we're saying here is that this technology could significantly increase your impact, and therefore, employees should invest the effort and time to go through that change. But this is not a change model we are used to. If we think about people at the top of the pyramid, such as senior underwriters, senior claims handlers, and senior brokers, they may not feel a need to change because they have reached a high point in their careers and believe they know the best way to do things. Every firm will need a strong change capability to get senior people to work in fundamentally different ways. You need tech talent, change talent, and people within the teams, such as junior underwriters and junior brokers, who are eager to embrace change. Combine that with strong role-modelling and a suitable nervousness of your competitors and you may get the traction you need.
What are some qualities insurance leaders need to bring to the team?
These qualities do not differ from what it takes to be a good leader in a volatile world in any sense. However, specifically to the insurance industry, leaders need to foster a culture where “inefficiency from reworking” is acceptable on the learning journey. This is different from being comfortable making mistakes. In insurance, making a wrong decision on a claim or when underwriting as part of embracing new ways of working is not a good outcome—we operate in a highly regulated business where we aspire to having repeat business with our clients.
If a team is using a new tool to improve our decision-making process, but after a week it isn’t proving effective because we had to override it with human decisions most of the time, a poor leader might say, "Okay, you've been less effective this week than you normally would have been. Let's stop using that tool. Throw it away and tell tech to pilot a new one.
In contrast, a good leader would recognize that the only way these tools will improve—and the only way we will get better at using them—is to accept that this pilot journey and R&D process will impact our efficiency as we get used to, train, and improve these tools. This is where classic leadership models come into play, requiring leaders to role model this behavior.
What may be slightly different is the greater degree of vulnerability when it comes to technology change. Typically, and this is a broad statement, the more junior people in the team are more comfortable with new tech tools and how to apply them. There are two main reasons for this. First, they are more often digital natives than their bosses and their bosses' bosses. Second, there is less downside for them. If you are an excellent energy underwriter, you might feel that the machine could help you make slightly better decisions on the margins, but overall, you are already one of the best. If you're an assistant underwriter or a junior broker, you know that you've got a huge way to go to become as good as your boss. So, you think, "Well, this tool makes me a little better, and therefore I want to embrace it."
This brings us back to the concept of the Digital Sherpa. If done properly, the Digital Sherpa helps junior people gain a disproportionate amount of experience. If they've only been in the job for three years, seen 100 claims, or underwritten 50 risks, the fact that the machine has seen 5,000 claims or 50,000 risks enables them, through the right prompting and nudging, to be much better. While the person at the top may only improve marginally, what you're doing is lifting those junior people.
How can talent be encouraged and developed to embrace these new technologies?
There are two parts to this answer. First, in a mid-sized insurance firm with 20 or 30 senior underwriters, you only need one or two to embrace new tools for the rest to then follow as they hear positive stories from their colleagues. If, say, the senior energy underwriter and senior property underwriter start using these tools and show improved KPIs, others will notice and want to adopt the tools as well. This creates a culture where everyone in their teams engages with the tools, and junior team members in other teams will start asking why they aren't using the same tools.
Second, looking at the executive level there used to be lots of hesitation when it came to tech. The situation has improved over the past 10-20 years. Previously, many executive teams lacked members comfortable with technology, making it hard to sell tech-driven ideas. Now, more CIOs, Chief Digital Officers, and even Chief Data Officers are part of executive teams, providing a tech-savvy voice. Additionally, COOs are increasingly comfortable with technology, and non-executive directors often have some tech experience, sometimes from roles in Insurtech startups or other industries. This shift helps bridge the gap between technologists and leadership, making it easier for executive teams to support plans to adopt and integrate new technologies.
My main message to executives is, don't fear startups or new competitors. The real threat is your current competitors who are adopting AI faster and more effectively than you. This means that the executive and non-executive teams don't need to be experts in the startup space or worry about external competitors from other industries. They just need to ensure that the culture and investment in their insurance carrier are ahead of their direct competitors. The incumbents that lose will be those that adapt slower than their peers.
What level of maturity do you currently see in the industry?
There is a low level of maturity. For big established players, being on the bleeding edge can be expensive and risky in a regulated environment. There is little appetite for starting projects that end up being abandoned, which is demoralizing for the team.
Insurance is a very transparent industry, especially on the commercial side, where people frequently share information about what's happening. It's easy to understand where your peers stand. There's an interesting conversation around being a "fast follower." This term is often misused to mean "wait and see," but it actually means being jogging slowly along and being prepared to sprint when needed. It doesn't mean sitting idle and then scrambling when others move ahead.
On the talent side, you need to keep your team engaged. Younger employees, in particular, may get frustrated if they see peers at other companies using advanced technologies that save time and improve efficiency. You need to ensure you're ready to move quickly when necessary and keep your junior talent engaged by demonstrating that your firm is appropriately engaged.
Having the right data is crucial. If your data is inadequate, you won't get the best results from your machine learning models. More insidiously, if your data seems fine during a pilot but isn't scalable, you face bigger problems. For example, if you pilot in one geography without considering the need to scale across all geographies, you might have to halt the project after significant investment.
Unlike many technology projects that can be done incrementally, AI initiatives often require substantial foundational work before scaling. It's like building a skyscraper: you need to dig deep and lay a solid foundation before construction. If you rush to scale without this foundation, the project may fail to deliver the expected benefits, leading to expensive and time-consuming rework.
Linking this back to people, you need team members who are willing to argue for the necessary foundational work. A brave CIO, COO, or CDO must advocate for a thorough approach, explaining that fixing data and building a solid foundation might take nine months, followed by a pilot and then scaling over 18 months. This long-term view is essential for sustainable success.
Who's doing this well?
Well, I think most people are doing this well in so far as they are approaching it slowly and steadily. What would be bad, in my opinion, is if someone decided to slow down their cloud transformation efforts because it's seen as boring infrastructure work, and instead, spent all their money on building the coolest AI-driven tool. That would be a mistake. Six to twelve months ago, many were talking about AI without really understanding what it meant. Now, most companies are aware that data is the challenge and that there's a risk in doing pilots before being ready to scale.
There is a need to discuss and gather use cases from across the business without overpromising. Most insurance companies, brokers, and carriers are probably in a sensible place now. From a technology evangelist's perspective, they might only be at one out of ten on their journey, but from a pragmatic leader's point of view, they're in a good place.
The companies that aren't in a good place are those that misunderstand what it means to be a fast follower. When the CIO or CDO says, "You need to build momentum now to be ready to run when needed," and they respond with, "No, fast follower means we do nothing until others have learned from their mistakes," those companies are at risk. In two or three years, they will have no foundation to rapidly move ahead.
What is one piece of advice you would offer to insurance industry leaders on taking full advantage of AI tools and implementation?
Encourage all your staff to use AI tools for at least 15 minutes a week to try and do they core job better. Until they start using them, AI remains just a concept. Once they use AI to help summarize a meeting or improve a Word document, they will begin to see its value. This hands-on experience is crucial for understanding and appreciating the benefits of AI and building the vision of how it can best add value.
