With AI continuing to upend and streamline the internal workings of the insurance world, data - and how it’s structured - is coming under scrutiny. Historically, data was harnessed manually, leaving room for human error and mistakes. Nowadays, AI is being heavily leaned upon to digitize unstructured data, reduce quote turnaround times and enhance the overall broker experience.
And it’s already changing the game, with Generative AI in insurance projected to grow from $1.09 billion in 2025 to a staggering $14.30 billion in 2034.
In a recent interview with Insurance Business, Katarina Pregelj, general manager, underwriter solutions at Applied Systems - which recently acquired Insurtech provider Cytora - revealed that digitization is in their DNA. According to Pregelj, their vision extends far beyond simply digitizing forms. For her team it’s all about creating an intelligent ecosystem where information moves effortlessly between agency management systems (AMS) and underwriting teams.
"We can decouple the method of placement from the need for that data to be structured and delivered to the carriers," she told IB. "That gives us a lot more opportunities to streamline the placement submission process across market segments and across different lines of business, irrespective of how complex that risk is.
"There’s a real opportunity here to utilize existing data, whether or not it's structured - and essentially use that data for the placement of new business or to automate remarketing or renewal placement. This will facilitate the flow between the agent and their markets in a much more efficient way."
But rather than agents driving every individual task, automation could increasingly manage routine workflows while people act as ‘AI managers’. As Pregelj told IB, at some point, she would argue that the human-in-the-loop will be managing different parts of the flow, rather than being the main driver of the overall process.
And the data agrees. According to recent studies, 80% of leaders say that human-in-the-loop AI is important for their business operations, with nine in 10 individuals adding that they’re more likely to trust a company that uses human-in-the-loop AI.
Aside from maintaining a human touch, one of the biggest pain points in commercial insurance remains duplicate entries between agency management systems and carrier portals. For Pregelj, she believes existing policy data provides a strong foundation for eliminating much of this work.
“For the portion of the existing business, we can utilize that data in the AMS to automatically trigger remarketing and essentially use AI and our orchestration workflows to move the data from one place, which is AMS, regardless of the form or shape of that data into the carrier system.”
“[For new business], Applied has a tool that can help collect the data in a digital way. It's streamlining that data intake as well as the placement in order to then get the submission to the right markets. At Applied, we want to help you assess what the right markets are - essentially doing that in a way that again decouples the method of placement while still making sure that all that data is under your scope and in your systems."
A large part of Applied’s Connected Trading vision revolves around Cytora and its AI capabilities, which are designed to transform unstructured submission data into information carriers can act on quickly and efficiently.
“Cytora is able to digitize any type of risk. And when I say any type, regardless of the complexity, regardless of what form the data comes in - it can be on a napkin - and make it adhere to the carrier's view of the risk. We’re able to either pull the data from the existing documents or also connect to third-party sources to really deliver everything that you need to make a decision."
One of Applied's latest developments introduces what’s described as a "stateful" submission process, allowing AI to continuously update a submission as new information arrives.
"What that means is that we know what your target state is and what you need to make decisions,” added Pregelj. “As different communications come in, we're able to map them and update the submission based on that new data. It sounds very basic, but it's actually technologically pretty hard to do - to have the stateful nature of communication and be able to track all of that and update it based on new information."
What’s more, this process means you don’t need to rely on the underwriter to tell the agent, 'Hey, I'm missing five data points.' Instead, it automatically feeds that back into the AMS. And, even more significantly, AI can determine whether those missing details already exist within the agency's systems.
"We will, given the agent approval, have an agentic process that looks into the AMS to understand, does that data point already exist, and if so send it directly to the carrier,” added Pregelj. "It takes away the time, it increases the responsiveness of the agency because now you almost instantaneously know what the underwriter needs in order to move forward. And increasing the speed to quote increases the likelihood of something being bound.”
Looking ahead, Pregelj told IB that the biggest challenges for the industry won’t necessarily be the tech, it’ll be removing the legacy systems professionals are all so used to and facilitating change management.
“Historically, in insurance, we gather the data in a form so that the other party can then remove the data from the form and put it wherever they need it. That has been embedded within insurance professionals for decades. [Nowadays], you don’t need a form - you don’t even need an email.”
However, this transition may be easier said than done. As organizations and leadership teams still rely on more manual, less AI-driven platforms and processes, this digital revolution won’t happen overnight.
“We're going to see certain segments, [such as] small commercial where there's already a thinner margin and more automated underwriting - so there's more incentive to streamline and optimize workflows.”
This article was created in partnership with Applied Systems
