Applications of GenAI in Insurance

Generative Artificial Intelligence (GenAI) is revolutionizing industries with its ability to understand, synthesize, and create language-based content. In this interview with AAIS Partner, BluePond.AI, we explored how this leap forward is particularly impactful in the insurance sector, where manual processes and complex documentation are prevalent. Hear from Pranav Pasricha, CEO of BluePond.AI, to learn what GenAI is, its transformative applications within the insurance industry, the challenges of adopting this technology, and how carriers can effectively navigate these obstacles to harness GenAI's full potential.

What is GenAI?

GenAI is a relatively new form of artificial intelligence and it represents the natural evolution in the path of AI, according to Pasricha. “What makes it special is its ability to actually understand language, be able to synthesize it, analyze it, compare it, and create sensible responses for you,” he said. “So, we see it as a big step forward, especially with the large language models because they now have the ability to ingest and analyze a lot of data in context.”

How GenAI Can Be Applied to the Insurance Industry

Pasricha feels that GenAI has a lot of applications across the whole insurance value chain. “We all know that the insurance industry still has a lot of manual processes,” he explained. “With GenAI, you can actually start to make sense of complex documents. You can read them like a human would, take meaning [and] values out, and start taking comparative steps.” For example, GenAI can assist with reading a claims note. It will look at what the claim is for and what the values are, then compare it to what's covered in the policy. From policy checking and comparing different quotes from carriers in reinsurance, to the ingestion of large treaty and portal documents, Pasricha believes the applications of GenAI in the insurance industry are endless.

Biggest Roadblocks to Utilizing GenAI

Pasricha sees five main roadblocks in the enterprise and insurance adoption of GenAI.

1. Security and Customization: Security and customization are important in GenAI applications. “When you’re using [GenAI] for sensitive and complex insurance tasks and doing it in the enterprise, it requires some sort of a secure environment,” Pasricha shared. “This involves a lot of integration and building custom models that understand insurance.”
2. Precision: There is no room for error in the insurance industry. “In this industry, a single error can cost a lot,” Pasricha warned. “If you're a broker and you didn't get the right coverage, or you failed to warn [the insured], that can be a major exposure. So, the risk of experimentation is very high.” Therefore, having a very finely tuned GenAI model and a good validation and QA (quality assurance) process is imperative.
3. Data Privacy and Confidentiality: Pasricha emphasized the significance of data security when implementing GenAI. “You can't just use a general open GenAI service because that will leak your confidential information, like your forms or your special underwriting rules, and become a part of the public office,” he stated.
4. Skill Set: According to Pasricha, expertise and modern approaches are essential for GenAI development. “You need to know what you're doing,” he stressed. “You cannot apply old knowledge, old methods, or old software techniques to building the GenAI; you'll just waste a lot of time and money.”
5. Time and Cost: Pasricha highlighted the financial pressures and urgency faced by organizations. “Nobody has endless budgets, yet executive teams and boards are asking for results today,” said Pasricha. “There is a way to get accelerated results from GenAI, but you need to be able to solve all these other problems to get it to work.”

How Carriers Can Best Navigate GenAI Roadblocks

Pasricha believes insurers need to have the right software foundation to be successful with GenAI. “You need to put in place a lot of building blocks so that you can build safely, securely, and with high accuracy and effectiveness,” he shared. With software like BluePond.AI, carriers have a partner that has already built some of those foundational building blocks. “Even if you want to build your own application or train your model with your own data, you still have a partner [in BluePond.AI] who knows what they're doing, has the right skill set and expertise, and more importantly, has an enterprise-grade software platform where you can test out all of your data privacy, security, [and] quality concerns to make sure that they're taken care of.”

About BluePond.AI

BluePond.AI is a relatively new firm with a team of more than 50 members, each bringing 15 to 20 years of experience in the insurance industry and a long track record of developing machine learning tools for the sector. “Our special sauce is that we have a platform that is very targeted and pre-trained for property and casualty (P&C) insurance,” Pasricha explained. “Think of it as a GenAI platform that, out of the box, understands complex P&C language, can identify P&C insurance documents, and can run very fine-grained operations and data extraction on those documents.” For example, for brokers, BluePond.AI can identify a policy, the line and class, and who the insured is. “In the context of the line of business, we can automatically do a policy check or a code comparison for underwriting at a carrier site,” said Pasricha. “We can look at compliance to underwriting rules for claims and match claims documents and policy documents for reinsurance, too.” BluePond.AI is a custom-built platform for the whole P&C value. Learn more at bluepond.ai.

To view the full interview with Pranav Pasricha, click on the video above.