Nationwide Spending $100M on AI to Beef up Claims Efficiency, Customer Experience

This article is part of an on-going series about technological and AI-related developments and advancements in claims processes and claims departments.View Series

When an insurance giant like Nationwide undertakes a revamp of its claims operation with artificial intelligence and new technology, there has to be some noteworthy things to say about it.

The Columbus, Ohio-based insurance and financial services company, which offers a range of products including auto, home, life, and business insurance, holds an A+ rating from both A.M. Best and S&P. The Fortune 100 company with more than 24,000 employees promotes itself as having a strong focus on customer service. The company began offering 24-hour claims service in 1960, reportedly the first carrier to do so.

Lately, Nationwide has been putting resources into AI and updating its claims operations. The company announced a $1.5 billion investment through 2028 in technology innovation initiatives, with $100 million earmarked for advancing AI each year for the next three years.

To get a lowdown on the changes, Claims Journal spoke with Guru Vasudeva, senior vice president and chief technology officer of property/casualty technology at Nationwide.

Vasudeva’s prior roles in the company were senior leadership positions including senior vice president and chief technology officer of infrastructure and operations. Before joining Nationwide, he spent six years at IBM as an executive architect.

The conversation has been edited for brevity and clarity.

Claims Journal: What new technology have you recently integrated into your claims department/process?

Vasudeva: In claims, we continue to really build on top of our investments in the Guidewire claim center. We recently also migrated all our contact centers to Genesis Cloud. That gives us better voice recognition, natural language processing and those big system implementations are already done. We continue to leverage the capabilities, both AI and non-AI related capabilities, that come from both claim center and the Genesis Cloud in the callback center. We also have done what I’m most proud of actually, our early implementations with the generative AI, we call it Claims Log Notes summarization.

Claims Journal: Why are you implementing this technology? What problem is it solving or will it solve?

Vasudeva: We get a lot of submissions as you can imagine, and we are using the traditional OCR technology. We are trying to move it towards generative AI-based technology for fast interpretation of the documents that are coming in and ingesting those documents, mapping them to the right fields and so on. It reduces manual work for our (claims service representatives) CSRs in the claims area.

When complex claims happen, people are dealing with a lot of stuff, right? I had a water leak a year ago in our kitchen. I remember that whole claims experience: Nationwide was incredibly great to work with. The reality, though, was it took a long time to work through all the moving parts. So first you work with adjuster, then you get couple of quotes, because it’s a big project. Not only is it a plumbing issue, now it’s a hardwood floor issue. In the process we realized that maybe we need to replace our fridge and so on. As a result, the whole experience, in reality, the implementation of the project might take months sometimes to really get through the whole claims experience. And even though Nationwide is incredibly responsive, the reality is you still have to work with all these different vendors to get things done.

So, when a customer calls the call center and asks for a status or asks for some information about that claim, the CSR has to read all these log notes from all of the various interactions that the customer has had with Nationwide and that takes time. Maybe it might take three-four minutes because you want to provide very comprehensive support and answers to the customer and you want to stand by what you say as a company, and that is a time consuming for the CSRs. It’s also frustrating for the customer because the customer has to wait until this CSR reads everything.

So, one of our first implementations we did almost a year ago that we put into production, we call it Claims Log Notes summarization. Since generative AI is really good at summarizing the information, we basically created an ongoing engine that says that ‘OK, Guru had a claim and he had an interaction.’ We recreate the summary from all of those interactions. So, every time a CSR looks up a customer, instantly there is a nice summary that says what has happened across all these log notes. That is an incredibly good use-case of using generative AI and we already put that in production. And it’s highly leveraged by our CSRs; they love the speed. It also helps customers because now they can get the answers quicker and they don’t have to wait for a long period.

The other benefit from all of this for us is that we are empowering our CSRs to really focus on the empathy aspect. We’re using generative AI with a human in the loop and that’s the key part. We’ve got another effort, which is Call Transcript Summary. There is a lot of after-survey call work up after call work happens in the call center, and we are leveraging transcripts after a call to pre-populate some of the information about the call.

Claims Journal: How is it working so far and how do you hope it works?

Vasudeva: We’re very proud of the mobile and web capabilities that we are already providing for personal lines customers for claims. As you may know, Keynova (Group) rates us number one in terms of claims capability and that is an intentional move on our part because we feel that customers really need the care; they need it at the time of when it matters most. We have implemented a lot of different capabilities. We’ve got an ability today for people after an accident to do submissions online, take photos. We’ve got estimation software that is also leveraging AI.

So, we are thinking constantly about how to make it easier for people to do the submission, follow through at the same time,have associates available to help them when they want to really talk to our adjusters or call center. So, what I would say is we’re always focused on resolution as soon as possible. Leveraging these technologies will really get us towards doing the claims adjudication as fast as possible, as accurately as (possible), so that we can really deliver on our promise.

Claims Journal: How long has it been up and running?

Vasudeva: Claims Log Notes summarization has been up and running for more than a year. The Call Transcript Summary was introduced a few months ago.

We are constantly monitoring because companies like Guidewire and Genesis Cloud, they’re coming up with some of these capabilities built into their software. Since we rely on these two big engines for our claims area, one for our claims system and the other one for our call center area, we’re constantly asking ‘Should we adopt what they provide out-of-the-box instead of leveraging something that we have implemented?’ We wouldn’t be offended if they had a better solution. We would want to leverage those kinds of things out-of-the-box. In fact, we do know Genesis particularly has got lots of capabilities around the whole call center experience, and that applies to claims as well.

The one area where we will continue to build our own solutions is how to make the entire experience seamless across multiple channels of interaction, the omnichannel experience. We have a phenomenal website and mobile app. The claims portion of it is ranked very highly by Keynova. But then we also have a call center. We also have people that go to the house and work with you as adjusters. So, what we are really looking at is how do we keep all these systems connected and empower all those channels so that we can serve our customers effectively.

Claims Journal: How do employees feel about the changes and the new technology?

Vasudeva: The Claims Log Notes summarization was the first production scale implementation we did; there was a lot of skepticism in the beginning. Our employees, they have a lot of pride in their work. They want to make sure that they’re really reading through everything and making sure that they’re answering the questions accurately. So, in the beginning, what we did is we…asked them to do what they have always done and also read the summary and start to report if the summary really gets them what they would have wanted…or can they start with a summary and then use the details when necessary, depending on the conversation where it went. We engaged the employees in this. This is not an expense-reduction play—it is more of a focus on customer experience and speed-to-answer. While there was some skepticism, we definitely have good adoption of this capability.

Claims Journal: How do you feel about it?

Vasudeva: I remember building our first 360-degree view of the customer. That project, I was the chief architect on it in 2009. We’ve been leveraging it to serve customers better using predictive models starting in 2010 time-frame. So, a predictive model, or what I would call traditional AI, has been around and we’ve been using it. We use it in so many different ways across the company that we are already good at.

Generative AI is completely different. The way I think about it is, (the) Internet made information democratic. In other words, it brought information to everybody. In fact, I remember reading that once we got to full scale adoption of the Internet, and then finally iPhone came along, pretty much every person who had those devices had similar information that the U.S. president did not have in their hands just 10 years before that. That is how profound that shift was in my personal take about generative AI—generative AI is going to democratize knowledge completely. It’s going to make it easy for everybody to get the insights that they need.

In the claim space, I think there is a lot of opportunity for us to leverage it—to do the adjudication faster. But we will always be focused on doing it with humans in the loop. And at the end of the day, when you have a loss, when you have an accident, that is a pretty serious life moment and we want to make sure that we are doing it with empathy and there is a human judgment involved in how we deal with our customers and claimants and we will always leverage these technologies. I’m excited about it as an aid to our claims associates. I’m excited that it will make their job easier and better so that they can actually focus more on talking to customers and helping solve problems for them.

Claims Journal: How will you measure the success of this technology? What are the metrics?

Vasudeva: In the call center portions of the claims, I will use the average handle time, the kind of metrics we are looking for those things to get better. The big outcome that we are looking for is, is it resulting in better customer experience? Are we getting after survey results? Were people really excited about the how they got served; that’s what we really look at. Our focus is more on those kinds of metrics and then of course those metrics change depending on the use case. So, for example, if I was giving you details about our claims estimation tool, then we would be looking for how accurate it is, right? And so on. So, it depends on the use case. Average handle time, first time, final, where possible, did they get the answer that they needed. And then last but not least is customers’ satisfaction.

Salazar is a Claims Journal intern. She’s a student at Cal State University Long Beach majoring in journalism. She expects to graduate in June, 2026.