How Insurers Are Harnessing Gen AI to Improve Risk Selection
Generative AI tools, such as Open AI’s Chat GPT and Google’s Gemini platform, are another weapon in the insurance industry arsenal if brokers and providers can get a handle on how and where they fit into day-to-day processes.
According to experts Gen AI will be able to help with a broad range of insurance applications. Alisha Fazo, Senior Director, Product Management, Moody's, noted : Generative AI represents a significant paradigm shift there. It's moving us into a new era. It's transforming the insurance industry by expanding beyond the conventional limits of our current applications today, and it's facilitating the creation of a comprehensive risk management ecosystem.”
Indeed, for some, the sky is the limit in terms of Gen AI related opportunities. As Scott Christian, Executive Vice President - Catastrophe Analytics, BMS Re, outlined: “It seems like there's always a bright individual out there somewhere building something creative that could move the needle and help the industry move forward.”
Crucially, the technology is capable of taking segmented risk management systems and data analytics platforms and powering them through one conversational interface. Insurers are already taking advantage of Gen AI to improve risk assessments, reconcile fragmented data, improve underwriting accuracy and contribute to loss prevention.
Fazo added: “Gen AI is enabling insurers to make faster, more accurate decisions in this complex and rapidly changing risk landscape.”
The technology is making it possible for insurers to understand data and develop key insights in a way that simply wasn’t possible before.
Oliver Mapp, Head of Data Science, Price Forbes Re, commented: “Using Generative AI now, it's possible to interrogate that data and lift insights on data in a way that it wasn't practical or scalable to do before. And you know, that's not just claims bordereaux. It is also true for policy bordereaux, statements of values, exposure databases, all of these data types that can flow through the industry.”
Essentially, it takes over the time-consuming, previously manual job of assessing unstructured data, but this doesn’t mean the technology is a cure-all, nor is it advanced enough to replace experienced insurance professionals.
“It's still early days for Gen AI, but a lot of the things are happening right now. I would call them efficiency play. It's kind of helping us be more efficient. It's not taking over entire processes or making decisions for us yet,” Nick Johnson-Hill, Vice President of the Data Group, Crum & Forster, explained.
However, opinion is split on how fast insurers should be embracing Gen AI. A poll taken during the webinar saw 45% of the audience suggest their firm was not progressive enough about using Gen AI. This is reflective of wider market sentiment. Fazo pointed out there are a lot of watchful waiters looking to see how other firms use the technology before jumping in with both feet.
As some of the market observe the impact of Gen AI, some insurers are already realising the benefits which are mainly focused around productivity and efficiency, but Mapp is certain these will translate to financial benefits soon: “We're going to see benefits in the expense ratio and the loss ratio. And I think that the expense ratio is going to come because the marginal cost of doing business is going to become less as we become more efficient, so we're going to be able to write premium at much lower fixed costs.
“I think the loss ratio [will improve], because it's just going to be possible to get across data and do a depth of analysis that its kind of just wasn't practical to do before.”
Claims could also be revolutionised with image-focused Gen AI. Johnson-Hill noted that you could have a photograph of a vehicle accident or roof damage and Gen AI can assess that damage and do something with that information. It could also be utilised for claims audits, comparing to best practice.
However, the models are used, users must have a clear end goal in mind. “It doesn't really matter which large language model that you select. You have a vector in place. It's more, ‘What is the task that you're trying to accomplish? What are you asking?’” said Fazo.
There are also risks anyone using Gen AI must consider. Gen AI is only as good as the information it can gather and there is the risk of ‘hallucinations’ - incorrect or misleading results that AI models generate.
Also front of mind for insurers is the ever-present spectre of regulation. They must navigate the complexity of different AI regulations in different jurisdictions – be that country or state - and monitor the changes which are evolving as the technology develops.
For Christian, one fundamental should never be forgotten by regulators or users of Gen AI. He insisted: “With respects to where we draw the line between what we want the AI agent to do and what we don't want it to do. I think rule number one should be; we rule the technology. Technology doesn't rule us.
“We still have models that are becoming more and more sophisticated…I go back to using these tools as companions, not competitors. But we need to make sure we don't over rely on the data they provide. Even being able to summarize a 100-page document in a fraction of a second and give you three bullet points, that's great, but I do think that we need to be cautious not to turn our brain off and stop thinking,” he cautioned.
As Johnson-Hill noted: “We need to keep a human in the loop, because there's risks here that we need to control.”
Whatever the risks, the genie is out of the bottle and Gen AI in insurance is here to stay. To succeed in making best use of it insurers must educate themselves. As Fazo detailed: “Start educating yourself. Start getting familiarized with it. And when you do this, and once we get it fully out there and matured, and everyone is integrated in some capacity to their solutions. You're going to see automation across your labor-intensive tasks.
“You're going to be generating new insights that you weren't able to before. You're going to enhance your overall decision-making processes. And with that, Gen AI is going to enable all of us risk management professionals in the insurance industry to be more efficient, to make more informed decisions, and to concentrate on the higher value activities where we want to be.”
For those who take the leap into Gen AI the future is bright and the opportunities are great. As Johnson-Hill concluded: “This isn't just hype. This is early days. It's getting better exponentially. This is the worst that AI will ever be, right. It's just going to keep getting better and better.”