Artboard 3 copy@2x

November 25, 2022

From the research team


Underwriting is arguably one of the foundational business functions that helped create our modern world. The term originated in the early days of Lloyd’s of London as people would sign their name under the amount of risk they would be taking on, and underwriting has since evolved into the cornerstone of insurance we know today. From those early days, underwriting has changed into a data driven function, and now this mature field is ready to partner with one of the youngest: Artificial Intelligence(AI). 

We are in the very early stages of AI implementation, but it is moving fast with new applications and AI models coming out almost weekly. Early adopters are already using AI for commercial applications, but most organizations haven’t yet tapped into AI’s potential. In the often slow moving world of Insurance this is absolutely the case, but it shouldn’t be.


The insurance industry is just starting to dip its toes into AIs capabilities with a primary focus on claims processing. On the claims front Lemonade’s AI powered lie detector got them in some hot water recently, but more conservative uses include estimating the costs of home contents based on square footage and predicting car repair costs off of telemetry data. 

Along with claims, insurers are also laying the groundwork for AI capabilities by moving towards API driven systems that can connect easily to models and AI. However, there’s not much focus on underwriting, which is a very natural use case for the technology.


Being excellent at underwriting is about understanding risks and their many forms, and risks come in a lot of unexpected ways. One example is the risk nuclear power plants face from jellyfish, unexpected but important. These types of risks are hard to know without considerable experience. One of AI’s strengths is its breadth of knowledge. Often trained on truly massive data sources, AI can use its knowledge base to make connections and inferences that humans simply cannot anticipate. 

When we pair underwriting and AI, we can augment the underwriters risk assessments with the vast knowledge base of AI. Ultimately, this means AI can augment the decision-making capabilities of underwriters and provide them with a treasure trove of information proactively—without requiring them to memorize everything.

This video script was written by the AI GPT-3, and the person speaking was generated by Synthesia AI.


Taking this idea one step further, one of the underappreciated commercial aspects of AI is its ability to be trained by best-in-class performers. Insurance today is losing a lot of knowledge from experienced underwriters who have specialized in niche fields, and AI is one way this knowledge can be preserved. With well trained AI, every underwriter can access the wisdom of a seasoned and specialized expert in whatever they’re looking to write, and insurers can double down on their underwriting success stories. In other words, with well trained AI your brand new underwriter can lean on the collected wisdom of your best underwriters to make far better decisions.


Underwriting has made our modern insurance possible, and underwriters today have a unique opportunity to truly shape the future of underwriting. Early adopters of AI will become absolute superheroes within their organizations, able to increase their efficiency and broaden their knowledge base at unprecedented rates. Along with the benefits of using AI, underwriters can also help train and correct AI mistakes, making AI better as it learns from the changes implemented.   

AI is a transformative technology that has the potential to revolutionize the insurance industry, and underwriting is an exciting area for this development. Predicting the future is always a challenge, but adding AI to the underwriting process is the start of a beautiful friendship.

Dais Research Report

Get weekly tech insights on important industry trends and stories. Stay informed and understand insurtech for free.