Get a demo
Select the type of demo you'd like to experience with Indico Data.
Select the type of demo you'd like to experience with Indico Data.
Watch Indico Data CEO Tom Wilde step in as co-host alongside Michelle Gouveia, VP at Sandbox Insurtech Ventures, in season 2 episode 3 of Unstructured Unlocked with Kelly Cusick, Managing Director at Deloitte.
Tom Wilde:
Welcome to Unstructured Unlocked, a podcast where listeners discover how insurers are entering the decision era, utilizing artificial intelligence to refine their decision-making processes, boost underwriting profitability, and achieve premium growth. I’m your host, Tom Wild,
Michelle Gouveia:
And I’m your host, Michelle Govea. Hey everyone. Welcome to another episode of Unstructured Unlocked. I’m co-host Michelle Govea.
TW:
And I’m Tom Wild.
MG:
And we are joined today by Kelly Cusick, the managing director at Deloitte Consulting. Kelly, welcome to the podcast today.
Kelly Cusick:
Hi, great to be here.
MG:
Thanks for joining us. So just to kick it off here, can you give us a little bit of detail about your background and your experience?
KC:
Yeah, sounds great. So by way of background, I’ve been working in the property and casualty insurance industry for over 25 years. I’m an actuary by background, so I always joke that means I’ve learned the industry from the ground up, meaning as an actuary you learn how the business of insurance works at a very detailed level and you have to think very granularly about all of the different financial levers and things like that. But at a certain point I wanted to broaden out beyond just the actuarial space, and I was fortunate to have a mentor with an underwriting and product management background, and so I worked closely with her to start our underwriting practice at Deloitte. And so I’ve really been focused in that space for the last 10 or 15 years, and I currently lead that business in the us.
TW:
Both my parents were actually actuaries. I always like to refer to actuaries I as the original data scientists.
KC:
That’s true. Yeah, definitely.
MG:
Well, we’re really excited to have you on today to talk about what you’re seeing in the space. I mean, you obviously connect with a number of the insurance carriers and have a pulse on what some of the biggest trends and themes are as it relates to needs in the underwriting space and obviously with AI and all of the new capabilities that come with that. Can you share a little bit about what you are seeing and what some of the, I’ll call it the bigger innovations or maybe the most exciting ideas are that you’re coming across?
KC:
Yeah, I would say in the underwriting space, automation has been a journey that the industry has been on for a long time and I like to talk about what I’ve seen a lot of the energy and investment in the past around has been, I guess I’d call it the bookends where it’s the beginning of the process and submission and how do we cut down on all the re-keying that we have to do and the messy process of the handoff from the agent over to the insurance company and trying to get quotes pulled together. Then the end of the process where it’s very much about the transactions and document management and how do we make that process less painful, sending things upstream and downstream and trying to track midterm endorsements and things like that if you’re in commercial lines. So there’s really been a lot of focus on that.
KC:
I think just because automation for a while was how do we just better build our foundational systems and more investments in modern core systems like policy administration and billing, and then things like robotic process automation and starting to use third party data. So that’s really where there’s been a lot of automation to date. I think what’s really exciting and some of the innovation and investment we’re seeing now is more in more sophistication in automation. Meaning how do we get not just process things better, but how do we get better insights and get better insights to our underwriters and those working in the underwriting space more quickly and more real time and AI is a big piece of that. And then generative AI in particular is a big piece of that because when we’re able to unlock the power of unstructured data, that’s where I think the underwriting, the tools that underwriters can use gets a lot more valuable. And so then you can start to hone in on that heart of underwriting within the process as opposed to the bookends of where they’re really going through the risk assessment and the deal making and making decisions and thinking about portfolios and how do they use their capacity and manage exposures and things like that.
TW:
In that middle part of the process you described, take us into the mind of the underwriter. I think that automation has been portrayed at times as a robot at times as a bionic arm. The way we like to describe it in my conversations with underwriters, I think they, they’re interested but quite wary, so of the giving up control or sort of black box decisioning. So take us into the mind of today’s underwriter in terms of their relationship with automation and AI right now.
KC:
Yeah, I do think there’s still a healthy amount of skepticism out there because underwriting isn’t art and you don’t want to diminish the human element of what underwriters do and just how they think about making decisions. And a lot of what they’re doing is they’re on the front lines of what we haven’t seen before. So yeah, there’s a bunch of history and they’re taking their personal experience and what they can see in the information to make judgment calls about how should I price this risk? Where should I set terms and conditions? Am I willing to write this at all kind of thing? Or if you’re looking at a portfolio, kind of setting the parameters for that portfolio, if you’re doing something like small business or personal lines where you’re maybe not touching every single risk. So in their minds there’s a lot of value that they bring to the table and there’s a lot of value and experience and having seen things, but at the same time, they’re also taking a world that’s very uncertain and making judgment calls on what the future might look like and using the tools that are available to them.
KC:
So that’s why I say I think there’s a healthy amount of skepticism, particularly when it was around things like, well, what can a model really tell me? And I don’t understand what data you’re using to put this model together in the first, so how can I trust it? But I think the opportunities that exist around automation and advanced analytics is number one, is around synthesizing information. So you as a human can only synthesize so much information at any one time. So if you have better tools and models that can do that for you, then it’s providing you guidance of, Hey, I did a bunch of manipulation and this is what came out. So it’s saving time from having to poke around a website and then go do a search and do a bunch of pivot tables on your own data and that sort of thing.
KC:
So can it do that for you? Take that heavy lift out of it. I think the other is the ability to identify predictive signals, particularly things that happen in combination. So I may know that if this is a restaurant that has a fryer that’s more of a risk, but if I have a bunch of restaurants that have fryers, what is that other thing or other factors that in combination make this a good risk versus a not so good risk. So again, the human brain can only process so much, so how can AI or some sort of models give me that. And then I think a third opportunity is around what I mentioned around portfolio management. Where can I have better tools that give me a pulse check on where things are at? It could be mix of business. So I thought if I’m managing a small commercial book, I was expecting certain proportions or targets of certain types of hazard level risk classes. And if I’m starting to see reporting coming out that’s saying like, oh no, we’re putting a lot more of this on my book. I can take action on that right away and try to understand why.
TW:
Or you’re saying for existing policies, understanding the complexion of the existing policy book
KC:
Or as new businesses coming in, am I starting to see a whole bunch of new business coming from a particular agent or geography or a risk class that I wasn’t expecting trying to understand why and do I really want this or not or in these days of property is certainly top of mind for a lot of companies. And so trying to understand they have a set capacity about what they can write and understanding risk concentrations. Some of it is risk management in understanding where do I have concentrations of risk in high hazard areas. But some of it’s more opportunistic where it’s, Hey, if I don’t have this data at my fingertips, I might cut off writing business in certain places sooner than I need to. And so I’m giving up some business that I could be taking on because I just don’t know where I’m at with my capacity at any point in time.
MG:
Historically, usually you’ve got some type of accumulation modeling that’s being run in the background or some type of guidance that is part of that. And I may be wrong in this, so please correct me if I am, but at least in my day-to-Day in venture space, we hear a lot about companies that are looking to be a solution as part of the underwriting workbench. So I think of all these things as an end-to-end set of tools that underwriters have, whether it’s checking how much, what their underwriting to affects that accumulation model, et cetera. Is what you’re proposing is that AI and automation can help streamline that in the sense that it becomes more readily available than maybe some of these lagging indicator analyses that are kind of more the standard? Or am I hearing that wrong?
KC:
I like what you said about lagging indicators, what we talk about a lot that traditionally a lot of the information that you would get as an underwriter or someone managing the business would be very much a lagging indicator or even if it was a good indicator of what I needed, you got it so late in the process, it took so long to put all the data together and get it disseminated out to you. So we see a lot of companies that are investing in their data infrastructure to be able to pull information together much more quickly and have it put out into dashboards as opposed to waiting for the big thick PDF to be sent out quarterly or monthly at best. So just trying to pull all that together. And then the modeling part is more of a value add of like, well, so what does this mean and providing, so it helps shortcut the, okay, I have to page through a bunch of stuff and then think through in my mind, what does that mean? It’s just providing a little bit of a quicker, Hey, you may want to go look at this part of the report. These are the things that flared up that were a change from the last time you looked at this.
MG:
I think what you’re talking about would be interesting too, is actually for new risks. And I think back to when cyber insurance, it’s still relatively new. Let’s speak in the grand scheme of things, but when it was brand new, like a brand new product in the market, there was such a push to gather as much information as possible. And typically that was through very, very long applications because they were asking tons of questions to try and get to what is the real risks that you have? And there were a number of components of a policy like ransomware, business interruption, and no one really knew what the big risk was. And in the cyberspace it changes a lot the type of risk. And so I can imagine that this more real time analysis of these are the data points you collected, this is what you priced off of that really was or was not an indicator of the actual risk. And so I imagine it’s a much faster way to reevaluate your underwriting criteria for new products as you’re trying to better understand the policy holder demographic, especially as you think through small commercial versus middle market, et cetera.
KC:
I mean, cyber’s a good example of this. Where can I start to tap into different data sources? That will help because some of the work that I’ve been involved with cyber as an actuary more from the actuarial side was normally you think about how do I come up with pricing for something based on past claims experience? But for cyber, first of all, for a while there wasn’t any claim experience or if was it was very limited. And then also to your point of things change so quickly, even if I looked at claims from five years ago, are those still the claims that are going to be coming in the future? So the ability to use different data sources where if we start to understand, okay, what drives claims is more about some of the internet traffic and who are you connected to and all that sort of stuff.
KC:
How can I start to develop partnerships where I’m leveraging that partnerships between the insurance company and their customers or maybe third parties in between to figure out how do I measure that risk in a different way? And I think you see this on the personal line side with the telematics and using telematics data because for a long time how you thought about underwriting and pricing, personal auto insurance was very much based on proxies for how you drive, where you drive when you drive. But if you can get actual more precise data, then you can start to hone in, refine more and more of the pricing for an individual based on where they really drive
TW:
In a similar vein, especially for commercial. But in things like workers’ comp, I know we’ve heard from customers who have said, we have these large pools of unstructured data typically documents in archives and we know there’s signal in there, but we don’t have an easy way to get that signal out and use it to feed actuarial model hypotheses. Right. And so is that something now, certainly with large language models becoming mainstream and useful, are you hearing that from your customers? Can we unlock this proprietary data that we’ve had for a long time but we haven’t been able to easily activate it?
KC:
Yeah, that is definitely a use case that I hear talked about. I would say I think that that is one where there’s a lot of potential value, but it might move relatively slowly because it’s still pretty difficult to pull together all the information that you need, the unstructured information that you need, and then it takes a while to build up the modeling capability based on that and then getting comfortable with it and that sort of thing. So I do think that that’s a great example of how you could pull together the synthesizing the information between what the underwriters see and think what the risk control loss engineers see and think what claims adjusters are seeing and hearing. And all of that is documented in PDFs in notes where everybody uses their own vernacular. But if you can start to use large language type models, text all that sort of stuff to pull those things together and then start to pull out the themes, I think that’s where you’re going to start to get some really interesting insights.
KC:
And because that’s based on your own internal data, then you start to get a little bit more of your proprietary insights. And that’s where I think you can get, so what the underwriter might see then is, okay, risks like this one based on what we’ve seen in the past, you should really be thinking about putting an exclusion in for this type of characteristic of the risk. Or here’s where you might want to set the attachment point or propose these self-insured retentions or deductibles or something like that or policy limits. That’s really interesting. But I do think that those sorts of things are going to take time just because it’s still a lot of disparate information that sits in a lot of different places. And you got to have the modeling that thinks through what that means and then figure out how do I get that in the hands of my underwriters in a way that they trusted and use it.
TW:
Do you think that technology has made this a new competitive arena or is it the same competitive arena just for the different set of tools? If you follow my question,
KC:
My initial instinct would be more the latter in that I think it just gives you better tools to do what you are already doing. What I’ve seen as far as what makes a really good underwriting organization is discipline. And I don’t think that that’s going to change. And I think where folks have gotten into trouble in the past is where it’s been very focused on just grow to grow, taking shortcuts or going outside of guidelines or going outside of what you’re trying to do from a risk quality perspective for various reasons. Going back to the art of underwriting, I think the principles around the art of underwriting and those who do it really well and have that discipline and really strong portfolio management skills and mindset are still going to be the winners. But I think that the developing tools that make that conducive is only going to accelerate the game for those that make those investments and have that mindset.
MG:
I think it’s really interesting what you said there, Kelly, because I think it’s easy to maybe mistake discipline for standardization in some cases. And the whole point of underwriting is that it isn’t art. And so discipline doesn’t mean that you react to every incoming submission that looks the same, same way. And the reason I wanted to double click on that is because we talk a lot on previous episodes about the talent gap in underwriting, right? And how will automation or AI maybe bridge that gap? Or to Tom’s point, how are these new tools going to be a differentiator for some of the folks that are in there willing to leverage them? So just would love to get your thoughts on where you predict the underwriting role to grow into in the next two years, five years, 10 years.
KC:
Yeah, I think about it two ways I think about it, both from the lens of an individual underwriter and how you might think about the skills that you would need to build and that sort of thing. And then also from the underwriting function perspective. So from an underwriting function perspective, I do think that there are new roles that have developed and will continue to evolve, and we call those tech trailblazers and data pioneers are like these net new roles that we’re seeing. So tech trailblazers, it would be someone that would sit within an underwriting function, but they’re really on the front lines of liaising with the data engineers, data scientists, actuaries, technology folks to understand enough of their world to know how you build software or how you build models or that sort of thing. But they still are grounded in the business of insurance and what underwriters are trying to do so that they can function as sometimes it’s product owners or something like that within the organization.
KC:
And they’re really driving what technology investments, what data and modeling investments are really going to add business value based on what our business strategy is and how that’s going to work. And then they also champion back to the underwriting organization, what’s going on? What does this really mean? How is it developed in a context that the frontline underwriters can understand and then trust and adopt? So that’s one. And then the data pioneers are kind of the same thing, but really digging deep into the modeling and data science and that sort of thing. Those are roles and different configurations that we see that are kind of net new, that maybe weren’t thought about as being part of an underwriting function traditionally. Then for the individual underwriters, I think there’s one sense of, okay, on the other side of the table, I need to invest a little bit of time in understanding what do these models do to a point so that I know how I can be using them and what the guidance is telling me and how to react to that.
KC:
And then I think the other thing is honing those traditional underwriting skills, meaning if I am not spending a lot of my day poking around websites and processing things and just pulling quotes together and all that, I have so much more time for negotiating, for deal making, for thinking about that portfolio management skill, thinking about what are the 10 risks that are coming that I need to be able to think about how to respond to. And so being able to then double down and really hone those skills, I think is going to start to become more and more of a differentiator when you’re not having to do so much of the mundane work every day.
TW:
Do you think of, there’s certainly some narrative that underwriters are as a population getting older and it will be difficult to replace them, you have this balance of institutional knowledge versus individual skill, and on the other hand, you have AI’s remarkable ability to process data at rates that far exceeds people, but people still far exceeding AI in terms of data synthesis or judgment. So how is the enterprise thinking about trying to characterize that individual skill into institutional knowledge? Is that something they’re actively working on, or do they accept that that balance has to remain in place? And if so, how do we feed new people, younger people into the role?
KC:
Yeah, I would say that when I talk to underwriting leaders, so chief underwriting officers or those that are leading underwriting teams, that is by far their number one concern. I mean profitability first of course, but when they just pick about running their teams, that is the number one thing because underwriting is traditionally an apprenticeship model where you learn from your manager, you learn from the seasoned underwriter that has been doing this for 20 or 30 years. So it’s definitely a concern. And the other concern that I have is the kind of feeding of the next generation. So how are we bringing in new folks that want to be underwriters and how do we attract them to the profession? And if they come in and they’re working on green screens and doing PowerPoint slides and Excel files, that’s not what they signed up for. So there’s kind of this push pull that everyone is dealing with.
KC:
And I think that some of the things we’ve been talking about around automation and all that, it actually kind of helps with both sides of it because one is if you can start to get more the data and some of the insights being generated more real time, but grounded in the way that underwriters think that claims adjusters think, that sort of thing, then that apprenticeship model can still work. And so you can take a shrinking population of seasoned underwriters and they can have more time and ability and better tools to help teach the younger generation. So if the younger generation, someone’s got a desk and they’re less seasoned underwriter and they’re getting information fed to them about risks like me, then there’s a little bit less time that they need to draw upon that seasoned underwriter. They really can go to them with really more complex things that are outside of the norm or just to validate the risk like me, why is this telling me this thing?
KC:
And then I think if there are underwriters that are involved in building a lot of these tools, that gives those more seasoned underwriters confidence that I understand more where this stuff is coming from. I can trust the first line of defense on some of that, and so then I can spend more of my time on the outliers or just teaching and honing a bigger population of people not dealing with so many referrals all the time kind of thing. So that’s one. And then I think if you look at the folks that are coming in that are more junior, then they start to see more interesting tools, more opportunities to get involved in things that are look like data science, and that is supporting the underwriting function in the insurance industry more broadly.
MG:
There was a period of time, this may still be the case, you definitely have more relevant conversations that are more on the pulse of this, but there was a time where the big innovation in underwriting would’ve been straight through processing. That was, you meet these criteria, we don’t even need an underwriter to touch it. We know this risk, we like this risk. And my perception is that that’s kind of taken a little bit of a backseat now to this AI and automation conversation where it’s not straight through processing per se, it’s just making what is presented to the underwriter more complete and easier to do that analysis on. And I don’t know if there’s a friction between those two objectives now or if maybe it’s one in the same, but what are you seeing? Is that still, and it could vary by line of insurance for sure, but just curious what you’re seeing on that front.
KC:
Yeah, I think that there’s still a lot. I guess I would say that the tolerance for true straight through processing for property and casualty only works for certain types of risks, certain lines of business, that sort of thing. A lot of times we start to hone in a little bit more on not direct to consumer, straight through processing, but more no touch, low touch, higher touch underwriting and start. It’s really a spectrum. And so when we’re doing things around underwriting transformation and automation, we really start with the conversation around how do you start to classify your archetypes of business into these different types of underwriting and what is it that drives a decision for it to be more low touch versus a higher touch? And when you’re touching it, why are you touching it? And then it starts to vary and sometimes it’s a customer segment, so small business tends to be lower touch just for various reasons.
KC:
And so then you start to get into conversations about, well, what do we want to kick out and why? And what part of the process and when do we want to auto decline versus not? And it’s a decision. Different companies have different philosophies and strategies around that, and then as you start to move up to more complex business, then you’re touching it more. But that’s not to say that there aren’t places where, Hey, I’m doing middle markets business and they may be fairly hazardous classes, but can I start doing more automated renewals, for example? But I want to look at the new business. So then that’s where I say being able to do straight through processing is a tool in your toolbox. I talked a lot about you want to have the infrastructure set up such that you can automate certain things and then you decide when and where you want to turn that on or off. Sometimes you make certain decisions about, well, I don’t want to have this fully able to be automated because that’s expensive. I don’t want to pay for all the third party data, or I don’t need to have this tool turned on for every single thing, that kind of thing.
TW:
Talk about the compliance implications of this maybe a little bit, or at least how you’re hearing this from your customers. I noted that the state of Connecticut, I think one of the first state DOIs, maybe department of insurance issued its AI guidelines now, or its AI requirements actually that they’re going to impose on insurers licensed in the state. And I often kind of talk about this or hear about this from customers as the data to decision chain of custody. So your ability to explain how a decision was reached, even though you may be using artificial intelligence somewhere in the middle of that, which historically, at least in the early days of ai, it was difficult to build AI that wasn’t a black box, it was by default, it was more of a black box. How are customers insurers thinking about delivering that explainability, for lack of a better word, to the state DOIs when they get these inquiries?
KC:
I would say it’s early days. I don’t know that there’s really enough experience with that yet to see trends. I think everybody’s trying to figure it out. So from the company lens, I think they’re just still trying to figure out what do we want to do and why in general? And then knowing, Hey, I’m going to have to explain this to a regulator. And so that’s why I think that they may not implement certain types of modeling in certain ways because they haven’t figured out yet. How are we going to make sure that we’re meeting all our compliance things, we’re just going to use this for ourselves internally. And then I think the regulators are trying to figure out how are we going to regulate this because they have the principles that they need to abide by, but then they also have the practical reality of we in the insurance department cannot hire legions of data scientists at this point in time, so how can we possibly understand it? And my hope, this is just my personal opinion, my hope is that we see good dialogue happening between the regulatory bodies and the insurance companies to try to find a solution that still is going to protect customers, protect solvency, all that kind of stuff, but also not hinder the insurance industry from doing what they need to do to serve their customers and have that social impact and be able to be the safety net when people need coverage.
MG:
The perfect segue to the next question I had for you, Kelly, which was thinking about the end customer, the insured, and all this, right? Given Gen AI has now made it so that everyone can access AI and just the trend we’ve seen over the past decade or so, I’ll call it the Amazon notification of everything, everything’s at your fingertips. You can get everything turned around in two days time, free shipping, everything exciting. Historically, the insurance industry is known to move slower for real valid reasons, right? Compliance, the regulatory landscape, large amounts of data that they have to sift through. What’s that balance that insurance carriers or underwriters have to find between the customer experience and still maintaining that level of diligence and compliance that is the standard across the industry?
KC:
It’s really an important role for the underwriters. They are the protectors of a lot of that, but they’re also the ones that can innovate around how can we cover this risk? I think one of the things that tends to get lost about the insurance industry is everybody likes to complain about their claims not being paid or their price of insurance going up, but there’s reasons behind that because the insurance industry exists for a reason and businesses would not be able to function if they didn’t have insurance. People would not be able to have cars and homes and all that and live their lives if they didn’t have insurance at that protection. And so part of the balance of having that protection is those that are on the front lines of issuing the policies, pricing the policies and handling the claims also have to protect the solvency of the insurance industry for the benefit of everyone.
KC:
And so there’s always that balancing act there. The one thing that I think is really potentially exciting and a huge opportunity around a lot of the advancements in data and technology and all that is the ability to not just be there when something bad happens, but the preventative part of the insurance industry. So working with customers to help them manage their personal risk or their business risks such that you can prevent claims from happening in the first place and understand what types of risks are on the horizon that you should be starting to think about or protect for, and doing that in ways that are more close to real time and more effective. Not that insurance companies don’t do that today, but it’s, it’s not exciting for us to think about insurance. It’s something you have to do every six months or every year, and it’s a slog and it’s not something that’s interesting.
KC:
So the more that you can meet business owners or individuals where they’re at and make it more engaging, where it’s like water damage is a huge problem in homeowners or in property insurance. And so if there’s ways to work with to understand when people are doing things to their home and making it more about like, Hey, this is a great enhancement you can make to your home, or a renovation you can make to your home, or something like that where it’s something that they’re thinking about anyway, and then the insurance company is there to help them with that and then help them understand how they’re going to prevent potentially million dollars, sorry, hundreds of thousands of dollars of damage to something. I think that that’s more compelling and a more compelling way that the insurance industry can start to use this data and technology and work with their customers more directly. Sorry, I was kind of rambling.
MG:
No, I mean, it is this constant push and pull of trying to ultimately an insurance company, the way they interact with their customer is through the claim side, but it’s also on the underwriting side in some capacity depending on how much automation you have in there. But those are really the two times and that you’re engaging with them and you want to make the process as frictionless, as seamless as possible because you want that positive experience, which may ultimately lead with the next interaction being at a not so positive experience just based on the catalyst for that interaction. So I think there’s just this constant goal of the insurance and carriers trying to educate a little bit of why the process is what it is, to your point, right? There is a real reason that things move at the pace that they do because the insurance industry serves everybody, and so you can’t jump on the first technology bandwagon that comes along. You do have to consider all of the interactions, all the touchpoints, all the data that you’ve got in house, and all the ramifications of doing something too quickly or not understanding what may be downstream if something isn’t executed properly. So I think it’s an important topic that is often top of mind for anyone, any of the big functions within the insurance carrier.
KC:
And I think there’s also this idea of partnerships and can the more connectedness of systems and data sharing help make those partnerships even more effective? Because insurance companies plus, I mean, there’s the distribution channel that we talk about a lot, getting to the customer through the agents and brokers because agents and brokers aren’t going away. They serve a purpose. So how do we make a lot of that easier? But also just the risk management types of vendors, so those that are helping with crisis management or safety management and things like that, or the construction industry and builders and contractors and that sort of thing. If there’s a way to make that ecosystem work together better, then I think you can also better enable what I was talking about around helping the customer more holistically manage their risk for their assets or their person.
TW:
Well, Kelly, thanks. This was a terrific conversation. We want to thank you for participating and the really interesting perspectives,
MG:
Kelly, it was really, really interesting to hear your take on what you’re hearing and seeing in the space, and I hope some of the listeners take away some advice from you on how they should be thinking about automation or trying out a new technology or capability internally. So thanks again for talking to us today. Appreciate
TW:
It. Yeah, thanks. Thanks, Kelly. Keek, managing director, Deloitte Consultings insurance practice. Thanks so much.
MG:
Thank you for joining us for this episode of Unstructured Online. You can find all of our episodes wherever you listen to podcasts today, Spotify, apple Podcasts, anywhere. Be sure to write a review if you like what you hear.
Check out the full Unstructured Unlocked podcast on your favorite platform, including:
