A conversation with Gautam Roy, CTO of Unum, at re:Invent 2023 | Amazon Web Services
- Hi, I'm Jayme Hart. I run business development for the insurance industry, and I'm here at re:Invent 2023. I have with me Gautam Roy from CTO at Unum. And we're here to talk about sort of the experiences here at re:Invent and the great work that Unum has been doing on AWS. So Gautam, can you tell me a little bit about your role at Unum, and what you're responsible for? - Sure, sure, Jayme, first of all, good afternoon.
This has been a great conference. The last couple of days was wonderful. So at Unum, so let me first talk about Unum, and then I will describe my role. So Unum, we sell insurance products and services. We protect 45 million lives worldwide, and we provide access, affordable access to employees for their disability, affordable access to employees for their disability, life, dental and vision, critical, endless products in an affordable manner. And this year has been great.
There are several awards and external recognition we got. Three of them I wanted to share with you and the audience. The first is been the most ethical company three years in the row, which we are very proud of.
We're in the list of the most innovative company by Fortune. So that's something that I'm extremely excited. - [Jayme] It's very nice. - And then the third is being a great place to work by U.S. news and reports. So about me, I'm the Chief Technology Officer of Unum, and we provide IT services under the office of the CTO.
It's three towers. The first is infrastructure, and that includes infrastructure from on-prem on the cloud. The second in this tower is IT operations. And this is providing 24/7, 365 nonstop operations. Not only for our internal applications hosted on our on-prem, but also with our hyperscaler partners, with our platform as a service providers.
The second tower is on shared services. This comprises data integration, APIs, workflow, robotics. And then the third is around application. And there are three segments on the application. One is call center applications customer facing, the second is corporate applications, and then the third is we are also responsible for digital application for our UK lines of business. - Right, excellent, so it's interesting.
People think of insurance as a very slow-moving industry with a lot of technology that is fairly stable over a long period of time. But Unum have really driven a lot of transformation. So how is Unum thinking about digital transformation going forward? - Yeah, so we've been on a digital transformation journey for the past five years, and the key word is transformation.
When we talk about transformation, it is really how we are thinking about it. The mind shift change, the cultural change, the change management part of it, in addition to the technology that we've been rolling out. And our digital strategy is to support our brand promise, enable our brand promise, and our brand promises with empathy, simplicity, and expertise, right? So we enable our brand promise, we support our business strategy, and we provide outstanding frictionless services for all digital interaction. - Very good, here at re:Invent, you see so much innovation going on across various industries, right, on AWS really kind of changing the way that companies are thinking about their business model, thinking about their technology stack.
What has Unum been able to do on AWS that you are most excited about? - Yeah, I would say there are several things we have done in the past five years. In fact, we started our journey with AWS as our first hyperscaler in 2019. So there are several big wins, ring the bell moments. I will share three that I'm very proud of in working with AWS and providing this services for our customers. The first is our call center transformation. So we had 26 call centers and we moved all those call centers to AWS Connect.
The project was initially scoped for three years. We got the project done in two years ahead of the schedule. So we migrated the legacy, the heritage called center technology, substandard processes from the on-prem platform substandard processes from the on-prem platform over to AWS Connect. And in doing so, it was not a like for a like change.
We added capabilities such as multimodal omnichannels. We added intelligent IVRs, we added self-services, we added workforce management capabilities. And these were done not just haphazardly. This was done purposely. This was done in conjunction with our business partners, our change management team, our product team, to ensure that the transformation we are driving supports the customer experience the best. So that's one of the program that I've been very proud of.
The second is, with AWS, is we are designing applications and deploying applications on the cloud, on the AWS cloud, using cloud native services, using serverless technologies, and these applications that we are developing. And we have developed our customer facing applications such as portals, such as mobile app for our personas such as client persona, consumer persona. And we have seen a tremendous increase in stability in functionality. Our CSAT scores have gone up. Even our employees are very happy to have this technology because they're getting less calls, less service requests as a result of that.
And then the third is the marketplace. We've been partnering with AWS now for quite many years. And so what we have done through AWS that the other technology services that we consume, we are now channeling them through AWS marketplace. - Very good, you mentioned that stakeholders ship with the business owners during that Amazon Connect implementation. We certainly see that as so critical with all of our technology initiatives here at AWS, that close relationship between business and technology is absolutely necessary in order to be successful. How did you convince your business stakeholders to come to the table on that type of change and how did they view the Amazon Connect project? - Yeah, so it's a great question.
So what we did, instead of first leading with the technology, we started the dialogue on the outcome. What's the problem we are solving? And we did it through data. We looked into our current processes. We did customer engagements, focus groups. We looked into our CSAT scores.
We looked into our touchpoint surveys, and we learned a lot as far as what's the baseline is, what are the frictions that are happening. We heard it from our customers. We saw it from the data, and then we received great feedback from our own operational staff employees from our own operational staff employees who are on this call centers, for example. So that was very good. So instead of, "Hey, here's a cool technology "we are going to move to."
We started off with what's the problem we are solving? Here's the baseline and here are the services and capabilities that we are going to release to the marketplace in order to get the required ROI from those implementations. - That's great, and I've certainly heard it suggested by your business executives that Unum now feels that it has a market leading capability in this space to self-service customers. - Exactly, yeah, before we were like seven during the business hours, and now our customers, both the client and the consumer via self-service capabilities like service over chat, chatbot, various technologies like web apps. So through these types of portal engagement, through this real-time SMS engagement, they can get services and the support they need throughout. And it doesn't have to be binded by the business hours.
- Very good, I know you've invested heavily in AI within this stack as well. How did you get your business users to be comfortable using newer technology like AI, especially for an insurance firm who's usually very shy of implementing new technology in the market? - Yeah, you are right. And historically, the insurance industry is risk averse. And for all good reasons, we protect 44 million lives.
We have responsibility in securing income at that time of their need. So we have to be careful in what we do, right? So when you talk about AI, I'm gonna switch because there's so much talk in this conference about gen AI, right? So let me talk on gen AI and it is sort of the same story on the AI, right? So what we did with the gen AI is it is not so much where we are leading it off with the technology. Our approach has been and will be human in the loop. As I shared with you our brand promise, empathy, simplicity, and expertise. Empathy is one of the virtue we don't want to lose.
Now, what technology and accelerators like gen AI, what they are doing and they will do for us is provide those accelerators, provide us those productivity boosts so that when we are interacting with our customers, we are providing them more richer data, more informed data, right data at the right place at the right time. - Absolutely, and I think that's an important point that you make, right? A lot of the use of AI and gen AI is not in terms of replacement of core processes, but really augmenting human interaction and making people capable of very quickly addressing the needs of their customer base. - Yeah, I was in one of the keynotes session and I still remember the phrase that was mentioned is really a mutual relationship between the data, human, and the technology.
And this really is. It is not like one is trumping over another. - Absolutely, on the point of data, we've certainly seen that as a critical component in getting your AI projects and gen AI projects launched into production. Can you talk a little bit about the data landscape at Unum and what you did to make it successful? - Yeah, so certainly, I will do that. Now, I should say this, otherwise my team would be very mad at me that the gen AI project, we were talking about gen AI.
So we rolled out a project, a gen AI project in production in a very record time in four months from understanding, playing with the gen AI technology to a full-fledged production rollout, supporting a business process. So we are extremely proud of the project when excellent. And very quickly, we are seeing the business outcome that we are hoping. Our AHC has lowered the accuracy of the answers. Our AHC has lowered the accuracy of the answers.
It's 91% accuracy. So it's all really trending in the right direction, right? Now, the data question is very important because this technology will not work if we don't have a good data source, right? Because it's sort of garbage in and garbage out. And that's always been the case and that's the case even today, right? We haven't overcome that obstacle. So for the data sources and just talking about this project, which we rolled it out in production, what we did and made sure that we understood where the data is.
We made sure that the data that we are going to reference other data that we need for the answers we are looking for. And what we did is when we looked into the data sources and scrubbed the data and the documents, we also worked with our business partners, we worked with our operational team to make sure that the documents that we are going to point to this LLM model are the right documents that we need to answer those questions. And in fact, these are the same documents, the same data sources that our operational staff, the same data sources that our operational staff, our client specialists are utilizing in answering the question.
What we are doing with this implementation, with this project, with gen AI, making it easier for our client specialist to answers those questions. - Yes, very good, and that's a good point that you bring up. It's not just your structured data.
It's very much your documents, your unstructured information that's critical for these tools to be successful. Very good, continuing on gen AI, what do you see as the biggest potential within Unum and even the insurance industry to use this technology to influence? - Yeah, definitely, at the call centers, right? As I shared with you, we have 26 call centers today at different aspect of the journey from administration, from onboarding, from enrollment, from claims. So we definitely see that opportunities that gen AI will bring in sort of, again, human in the loop to make those conversation more personal, make those conversation more accurate to provide a quicker, richer feedback to the customer. So that's one area. The second where I see where gen AI would help is in the RFP proposal process. We have to reply to a ton of RFPs.
So getting that optimize, making sure we are responding to the right RFPs with the right information. So that would be helpful. And for the down market, I think where gen AI could help would be from the digital coding perspective. Those are the three big areas of opportunities. And then internally, apart from the business operations project, I think, is how we do our productivity, writing documents, sharing notes through teams meeting. That would be a second area where gen AI I see at Unum that would take on.
And then the third is on our development, right? So writing faster, better code, and rolling them to the production. So getting applications, releasing applications with less defect, with less investments. That would be a third area where I see gen AI could play a role. - Absolutely, and all productivity enhancements, right? - [Gautam] All productivity enhancements. - The RFP process, the summarization to drive sort of research capabilities, those sound like big business problems and that's the CEO or a board level conversation. - Definitely, definitely, and in fact, the project that we rolled out that was demoed to our senior leaders, that was demoed to the board, that was demoed internally, and as I said, it takes a village to get it done.
It's not just a technology project. It was a technology business and a change management. - Excellent, well, I have one last question for you here today, Gautam. - [Gautam] Sure.
- The AWS cloud journey is often a long and transformative journey. Where do you feel you are today on that cloud adoption journey? Are you at the beginning, in the middle, somewhere near the end? - Yeah, definitely not at the end, right? I think we are in the middle. We have done quite a lot of work in the last five and a half years with hyperscalers and definitely with the AWS. There are three areas where we are working and we would do more. We want to do more.
The first is mainframe optimization, right? So we are trying to understand where would we run the workload on our mainframe, on-prem, hybrid, in the cloud? So we are trying to figure that one out, what we want to do with mainframe optimization. So that's one area. The second, we have a cloud smart strategy. And with that strategy, we are reducing the footprint of our on-prem data centers. So we are pushing more and more services to the cloud.
So what we need to do is to continue that journey of moving, again, cloud smart. Not everything is meant for the cloud to identify, continue the process, and reduce our on-prem data centers continue the process, and reduce our on-prem data centers to really micro data centers, right? And then the third area that we are planning to work with AWS on our journey to the cloud is on our data, right? We have started the work on data of moving our operational data stores, of moving our enterprise data warehouse to the cloud. We are exploring what to do with MDM, so we wanna make sure that our modern data platform is that we progressively move towards the cloud, right? And then the last one I would say to wrap this round is to have a very robust FinOps and security structure is to have a very robust FinOps and security structure and a governance architecture around it to make sure we are doing it in a right way. - Absolutely, and it's certainly our goal to help you here at AWS with that journey and to create something that is just as easy as a smartphone to start to integrate your applications and run and manage your business in the future. So I look forward to the continued partnership, and thank you for sharing today. - Thank you, thank you, Jayme.
Thank you, Jayme.
2024-03-13 16:27
