(Image credit: Priscilla Du Preez/Unsplash.)

I’ve worked with hundreds of insurance companies over the last two decades, and the industry culture tends to take a “people over process” approach, by which I mean many companies avoid technology investments that are purely about automating people out of jobs. There are obviously many reasons that more issuance, underwriting, actuarial, and claims adjudication tasks haven’t been automated (legacy technologies, agent relationships, complexity) but a culture of maintaining humans in the process is a major explanation given by insurers themselves.

While this is not universal—even insurers who put their people above any technology strategy still take opportunities to automate lower-value areas—this current runs strong throughout the industry, even within large multinational insurers heavily invested in innovation. It has significantly influenced the evolution of insurance IT.

This culture has its downsides. It can mean some less efficient processes stay in place that an insurer has avoided automating to protect jobs, and this can feel like putting one’s head in the sand rather than accepting technological progress. I’ve seen wonderful regional insurers get acquired by national players and are then mostly shut down or distributed to other locations, and while I don’t blame this on that regional insurer’s lack of IT optimization, it’s at least a contributing factor to that end result.

The culture also has its obvious upsides: so many insurers really do put their people first. This results in insurers that become the heart of their local community and economy, insurers with opportunities for advancement no matter where someone starts in the organization, and insurers who put money towards retraining rather than layoffs. There are still so many examples of people working at one insurer for their entire career when that path is such a rarity in other industries.

A common mistake made by many insurance industry technology vendors (mostly startups, but even established players) is presenting automation as their central message. “Think about how much money you’ll save by cutting jobs,” is an easy way to lose the room and end the sales pitch on the first slide. Instead, the proper approach is to talk about augmentation, and how technology can help get more value from existing people and how it can allow them to focus on higher-value tasks. (“Think about how much growth you can get without having to hire more,” is a much safer message, especially in the remote work era where insurers are struggling to hire.) Insurance executives aren’t fools, and they understand that full automation is one possibility, but it’s best to leave that discussion for internal meetings rather than pushed by a vendor salesperson.

Does Generative AI Change This?

Many people, both inside and outside of the industry, are thinking about how Generative AI will automate tasks for insurers. I’m one of those people, and I’ve been talking about and writing about applied AI use cases in the industry for many years. But ruminating on all the use cases ripe for automation is putting the cart before the horse, or at least it’s only asking half of the question. The other—very big—half of the question is: will Generative AI (and AI in general) change the insurance industry’s culture to be more aggressive in automating processes, even if it puts jobs at risk?

There are a few reasons to think that, yes, this latest evolution in AI technology will change things.

1. The easier it is, the harder it is to ignore. Generative AI and subsequent generations of AI technology will simplify the automation of processes that have traditionally relied on human intervention. As automation becomes simpler, insurers will find it increasingly challenging to resist its adoption. While insurers might be willing to forgo spending hundreds of thousands of dollars to automate a process that pays off in five to ten years, the scenario changes when they’re confronted with automating something that promises immediate benefits for a smaller investment, regardless of a culture of putting people ahead of technology.
2. The better it is, the harder it is to ignore. There is real evidence to show that now or in the near future, these technologies can produce better results than previous generations of technology and better results than human effort. Many past automation efforts simply shifted the burden of human effort from the task itself to reviewing the results of the automated task and correcting errors. If new generations of AI technology end up improving the result rather than just reducing time/cost, it becomes difficult to justify keeping humans on those tasks.
3. Hype Means Investment. When everyone in the world is talking about it, there is pressure to do it. Boards start asking the CEO, the CEO starts asking the CIO, and expectations are set that begin pushing through cultural resistance. This isn’t just peer pressure: if your competitors are investing in something and you aren’t, it can result in adverse selection as consumers take the smoothest path to a policy. While hype doesn’t always result in significant long-term insurance industry investments (see: Blockchain), if the value is also there, hype + value is a compelling force.

There are also reasons to think that, wait, this isn’t something new and won’t bring about major change:

1. Culture! This is not the first time we’ve seen the culture of the insurance industry stand firm in the face of overwhelming technological evolution. It’s 2023 and many insurers still don’t have compelling self-service portals. If many insurance companies can stay the course in the face of the Internet, then they can weather this AI storm as well. I’m exaggerating, obviously, and even insurers without fully digital self-service still have been significantly influenced by the Internet in other ways, but that’s exactly the point of culture trumping technology. Insurers will integrate Generative AI within a framework that complements their overarching business strategy, rather than investing in a manner that upends everything.
2. Hype Means Hype. It’s not yet clear that this technology will change everything for insurers, forcing them to adjust culture and adopt it. Is this truly transformational technology or just iterative steps towards better and easier? It’s possible the effort-to-value calculation is reduced but isn’t eliminated. Given the current market conditions, numerous insurers struggle with hiring IT staff in general, let alone AI experts. So even if the technology’s implementation cost is manageable, the challenge might lie in recruiting (and paying the salaries of) the necessary talent. The point is, we can all expect some positive results from AI, but that doesn’t necessarily mean a radical shift or that getting there will be cheap. It’s likely somewhere in the middle. And for some insurers, that allows for delays as they defer to their existing 5-year road maps.
3. Is AI producing real bottom-line value? After five plus years of PE and VC firms pouring billions of dollars into insurance industry startups, how many startup carriers have actually beat out the incumbents and become the big industry players? Most of those startup carriers are leveraging every technological advantage they can manage, but at best one or two of them have captured a noticeable fraction of sales in their line of business. Realistically, most of the successful startup carriers will be acquired by existing large insurers and folded into a more traditional way of doing things. This isn’t to say that technological innovation isn’t critical to the insurance industry. But the basic model of insurance profitability relies on the exchange of capital for risk, and while technology can improve and optimize that model, it doesn’t usually upend the dominant players.

The Future of AI in Insurance

The insurance industry has always been a paradox in terms of technology adoption. It’s one of the first industries to truly embrace the modern concept of corporate IT, the legacy of which is still running today in many data centers and slowing the industry down. It’s an industry that understood the importance of data long before others, driven by actuaries and underwriters always looking for the next key variable, yet still struggles with its data infrastructure. And it’s an industry that has indeed seen major business model disruption from technological innovation while managing to continue to run stubbornly beside that disruptive model with the traditional approach.

There are certainly use cases in the insurance industry where Generative AI automation can provide better results that are easier to implement than previous waves of technology. But “better results and easier to implement” is not the same as “radically better results and cheap to implement.” To break through cultural resistance, it needs to be the latter. Most insurers have always taken a wait-and-see, slow-follow approach when it comes to technology, and this will be no different. Generative AI and AI in general will indeed make its way into the insurance industry, especially as it evolves to be better and easier. But just as insurers will adapt to new technologies, new technologies will also adapt to fit insurers.

To talk more about how generative AI might be used for real but practical change at your organization, reach out to me on LinkedIn.

As Data’s Importance Increases, Data Itself Becomes the Market