Generative AI reaches insurers in three waves: efficiency, automation of unstructured data, and AI-powered insights for decision making. To reach the third wave, insurers should start with specific business problems and adopt a data-driven mindset.
Who This Is For
Carrier and broker executives — including underwriting, claims, operations, and data leaders — who are moving past generative AI pilots and want to know what it takes to use AI for automation and decision making across the insurance value chain
In Brief
- Efficiency is generative AI’s first wave. Automation that interprets unstructured data is the second. AI-powered insights for decision making are the third and most effective.
- Insurers investing in advanced analytics and AI are already seeing returns. In WTW’s 2026 survey, leaders were more profitable than slower adopters, with combined ratios 6 percentage points lower from 2022 to 2024.
- Decision-making AI depends on data, and data quality and accessibility remain among the biggest barriers reported by insurers.
- Start with a specific business problem, then build a data-driven mindset, clean up your data, and put governance in place before you scale.
Whether it’s writing emails, summarizing meeting notes, or creating documents, insurers are already beginning to realize the efficiency potential of generative AI’s “first wave.” But preparing the insurance value chain to apply the next two waves — AI-powered automation and decision making — will first require some homework.
Generative AI is advancing how insurance companies think about an already evolving, digital value chain. While insurance carriers leverage AI to automate processes and deliver more customer services digitally, the insurance value chain remains a gold mine of optimization opportunities.
Like most industries, insurers recognize that generative AI and other large language model (LLM) technologies represent great potential for increasing efficiency. More than half of the North American property and casualty insurers in the WTW 2026 Advanced Analytics and AI Survey already use LLMs and generative AI, and another 29 percent plan to adopt them within two years. From generating customer-geared emails to summarizing meeting notes, preparing presentations, and building bots for routine customer questions, AI can eliminate a lot of time and drudgery from workers’ lives.
Improve the Insurance Value Chain With AI
However, efficiency represents only the first wave of generative AI. Think of it as the “low-hanging fruit.” In the second wave, AI will join forces with other digital tools to systematically drive more costs out of the value chain while increasing accuracy, efficiency, and customer satisfaction.
Consider the claims process. Many claimants can now submit claims more easily using an app or website, but evaluating the claims and processing them still require human intervention. Claims also come with other documents like police reports or medical records that insurers must also process. The data on those documents can appear in various places or even be handwritten, adding additional challenges to processing. Only 14 percent of insurers in the WTW survey use straight-through processing to automate claims workflows today, though another 36 percent plan to introduce it.
In the second wave of generative AI, proven solutions like robotic process automation and machine learning will combine with AI to interpret and evaluate such “unstructured data.” The combination — known as hyperautomation — moves AI beyond simple efficiency into the realm of understanding. For example, with one client, we’ve demonstrated that AI can learn to recognize and process similar types of data on differently structured forms. For another, we created AI that can turn mathematical expressions into human speech in multiple languages.
Generative AI for Insurance Company Decision Making
It will take time and large investments to mature these second-wave AI technologies so they can scale to meet the needs of a large, diverse industry. However, when insurers realize the potential, they’ll likely want to start adopting them.
Carriers and brokers that advance through the first two waves will enter the third and most effective wave of AI: generative, AI-powered insights and analysis.
In this wave, insurers will employ generative AI to examine the insurance value chain and, ideally, enable valuable insights that capitalize on interconnected data across the enterprise. The third wave of generative AI involves an interactive approach to being data-driven, providing leaders with relevant information when it’s time to make key decisions.
Tangible examples of this wave are still emerging, but insurers are moving toward it. Only 16 percent of insurers in the WTW survey currently use AI to augment human underwriting, yet 60 percent plan to prioritize it by 2028. The potential is clear.
Consider an underwriting leader with relevant trends at their fingertips as they contemplate changes in appetite. They can evaluate market trends, internal loss, and submission data instantly — and in a consumable way — by interacting with a model tailored for that exact purpose. Similarly, a claims manager responding to a catastrophe who needs to quickly understand the size and scope of the company’s exposure could use a curated generative AI assistant to understand the extent of the loss.
Prepare for AI-Driven Decision Making
So, how can carriers and brokers gracefully navigate the waves of generative AI?
Our advice is to step back and start with the basics: your specific business problems. At the highest level, insurers are quite similar. They sell products to customers, collect premiums, invest premiums, and pay out claims. However, each insurance company’s problems will be unique. Maybe you serve a large population of people who speak languages other than English or Spanish. Maybe you’re in a part of the U.S. that’s subject to recurring catastrophic events. Maybe your workforce resists change. The possibilities are unlimited.
Once you’ve identified your business problem, ensure your leadership adopts a data-driven mindset. Being data-driven means:
- Leading with a data strategy
- Basing decisions on the value of data, not the cost of technology
- Expanding decision making beyond the IT department
- Knowing what data has the most potential before creating storage or processing solutions that seek to address all your data
- Identifying the most important reports and metrics you hope to gain from your data
- Being open to building on small victories rather than overnight, transformative change in your data strategy
Conclusion
Adopting a data-driven mindset is critical. It paves the way for next steps in preparing your data for generative AI, like identifying and either removing or reprioritizing redundant, obsolete, and trivial data. Insurers already feel this pressure. In the WTW survey, 42 percent of respondents said significant barriers included data-related issues like poor quality and limited accessibility, as well as inadequate IT support.
Then, you must create a data governance strategy for using your data safely with AI. For insurers, that strategy also needs to meet regulators’ expectations. Many states have adopted the NAIC’s model bulletin on insurers’ use of AI systems, which expects insurers to maintain a written program for using AI responsibly.
These steps will help prepare you for the next waves of AI and whatever future waves may bring.
Are you struggling with insurtech overload? Our insurance team is uniquely positioned to provide a combination of pragmatic solutions and services focused on achieving your objectives. We’ll put it all together so you can avoid common pitfalls like disjointed, unconnected solutions. Contact us.