The insurance industry is still largely reliant on manual processes, particularly when it comes to lines of business that use agents and brokers for distribution.
However, with the power of today’s rapidly maturing AI technologies, how can you seamlessly integrate your data with data from big tech platforms to make better, faster underwriting decisions?
In this free industry webinar join a panel of experts from Accenture and Amazon Web Services as they tackle this key question and dive into how AI has emerged as the transformative technology in the insurance industry.
Watch the free webinar today and gain insights on:
Paul: [00:00:00] Welcome to today's webinar with Accenture examining how Artificial intelligence Is transforming Insurance underwriting. I'm Paul Lucas, the global editor of Insurance Business. I'll be your host today for a webinar that could help you make faster and better underwriting decisions. Now, before we begin in earnest, a few housekeeping notes to run through. If you do need any technical support during this webinar, please use the chat box. We do have a team on hand to help you with any issues. And finally, this webinar rekeying will be made available to all attendees at this event. So if you do have any distractions during the live events, don't worry, you will get another opportunity to view this. Anyway, back to the subject matter at hand and ask yourself this question How much do you really know about artificial intelligence and how it applies to the insurance sector? AI technologies, which are rapidly maturing, hold the key to integrating data and breaking with centuries old paper based decisions as insurers increasingly demand digital options. Well, this webinar, sponsored by Accenture will take you away from that superficial nod towards all things digital and put you on track for a new, more efficient and yes, more profitable era. It will examine how I can help insurance carriers gain faster access to revenues, how incorporating AI into underwriting can reduce the time spent on manual tasks, how the experience and satisfaction of brokers, agents and customers alike can be improved. And above all, how AI is driving value for insurers. Now, at the end of this presentation there will be a question and answer period, so be sure to type any questions you have into the corresponding Q&A box that's just on the right hand side of your screen there. And remember, the more feedback we get, the more we know about the issues you're facing around AI. So the more opportunity that we have to get you on the right track. So please do engage with us today by using the Q&A function. With all that said, to show us how to capitalize on all of these opportunities, I'm delighted to welcome Corey Barker, Michael Reilly and Aarti Gupta. Now, Corey is the managing director of insurance at Accenture with deep experience in the Sector. Michael is principal director of insurance at Accenture with more than 20 years of experience, while Aarti is insurance industry, head of business development at Amazon Web Services, where she is responsible for go to market strategies and implementations with insurance account teams. Now they are going to help you make the most of this opportunity, which is no longer as futuristic as it seems. It is very much here and now. Corey, Michael, Aarti, over to you.
Aarti: [00:02:48] Thank you. Can you guys hear me okay? Just want to make sure. Okay, awesome. So thank you so much for having us here and looking forward to the discussion today. I wanted to actually start back a little bit, take a step back. I think the topic at hand is of great importance to all the insurance companies, and we definitely see that even from an Accenture partnership standpoint. But like I said, taking a step back, what you really see on the screen, I wanted to kind of just highlight the six big areas of focus from a technology standpoint, from an innovation and transformation standpoint in insurance that as AWS as Accenture are engaging with insurance in all of these areas. And what that really means is what you see on the left hand side, right? So when we talk about the core systems of record and insurance, really the heart of the insurance companies, your policy system claims underwriting, billing, so on and so forth. So those have really fallen more into what I like to call transform the data center, right? So when insurance companies come to us in particular, they are really kind of talking about how do they move out of that legacy infrastructure into the cloud to really kind of retire their technical debt and then get faster and more agile, more innovative and speed to market is exponentially better for them once they have kind of gone through that journey. And what you really see on this slide on the right hand side is where the insurance companies like yourself are wanting to really tap into the topic at hand today. How do they leverage data? How do they kind of enrich that with AI ML models on top of that and really kind of use the digital capabilities. And I don't just mean having a mobile app or a D to C capability, but it's more about mining the interactions and then leveraging that data to come up with better customer insights. So for example, some of the areas that we will touch into, especially as we kind of talk about underwriting claims, automation are the areas around how do you kind of incorporate that data that I was talking about, the data analytics AI ML. So some examples would be called center transformation, how you use chat bots and placing personalization engines behind those direct portals and call centers to bubble up the next best action, next best offer. And then not to miss that the incorporation of data lakes. And I think a big part of our discussion is going to be kind of focused on that too. How do you incorporate data lakes into your big data estate? Right, Because that really becomes a critical component for helping to both liberate the internal that structured data that insurance companies are really good at leveraging from the legacy systems, but really to enrich that data with external and unstructured data sources. Because once that data resides in the data lake, it becomes an accessible target for AI and ML type of services, which we're going to be discussing more deeply today. And then here, the point I really want you to take away, right? I mean, this is kind of broken by purpose in three different areas. What we are seeing that the the most of the innovation, most of the transformation within insurance is really happening at the confluence of digital capabilities, data analytics, AI ML So that top section, the front, the middle section, middle office and then the back office, how do you kind of really take those and then kind of use that in a very complementary fashion. So the more frequent, the more digital interactions that your organization and insurance company will have, you can take that data mine and analyze that for greater customer insights, which can be used to better price risk, to support growth, retention and so on and so forth. So AWS services and again, this is just like a little kind of to show the breadth of the different services from an AWS capability which have been used for the solution we are going to be talking about in underwriting. It provides that single chassis of services to enable those core capabilities across front, middle and back. And this is critical as insurance companies, when they come and talk to us, they tell us that they not only want to have digitization and automation of the processes, but how can they leverage more of those digital interactions to capture and analyze and then use those for better decision making, so on and so forth. And some of those examples could be in your call center, for example, using the data for sentiment analysis or extracting information out of your underwriting documents or claims documents to automatically process some of the information that your underwriter typically would do or auto adjusting damages to vehicles and so on and so forth. And moving along. One point I would like to highlight here that, you know, the information that you see today and then the you know, the possibilities from the joint partnership that we have between the two organizations, between Accenture and AWS, is really coming top down. So both the CEOs of our companies, Julie Sweet and Adam, they have kind of come together and put a lot of energy and resources into kind of putting together innovative solutions in the insurance industry that your organizations can take and then quickly deploy and run with that. And then the last thing I would like to say that again, you know, I've worked in the insurance industry for over 22 years. One of the biggest challenges we had was there was a lot of data. Insurance companies tend to be very, very data rich companies. However, the amount of data has really exploded. And I think we can all agree that the video images, emails, you name it, phone calls, all that semi-structured and unstructured data, how do you really use that to enrich your structured data? And then how do you use the AI ML models on top of that to really kind of enhance the decision making? And that's what we really want to kind of talk to you about in the space of underwriting. So with that said, I would like to turn it over to Michael. We'll walk you through a solution that Accenture has put forward on that front. Michael, over to you.
Michael: [00:09:48] Yeah. Thank you so much. And guys, I'm excited to be here today and look in insurance. We have been talking about AI for a long time, right? You guys wrote a webinar at a conference five years ago. I was probably a hot topic. But at the time, what we've seen is most of our uses and most of our deployments to date have been tiny little pilots or tiny little experiments here and experiment there or whatever else. But we never really transformed the true business kind of from end to end. And what we're really seeing now is we're on the cusp and actually we're beyond the cusp of doing that transformation. We're actually seeing the value of artificial intelligence and the data and analytic technologies that go with it, transforming entire processes and processes that have been stuck in the same way for quite a while. And the reason for this is obvious, right? So if we start on the right hand side. We've known what the challenges are that we had with the data as where it was today and the pressures that are coming to the business to make us want to change in commercial underwriting. We're seeing our distribution. We're seeing either the brokers merge together, which is giving them more power, so they're demanding more quoting from our underwriters. Right. So we see that consolidation or we see explosion of new sources where I have to meet the market where it is. And some of those are digital channels and some of those are Web channels, but they're expecting me to provide information, insights at a much faster than I have today. The work has gotten more complex. In order to drive cost and efficiency, we're spreading work in different areas. We're trying to do it in different pieces. And the risks that we're evaluating are not the simple risk they were before. Right? Everybody has multiple paths. Everybody has multiple elements going on, and that is creating a lot of complexity. At the same time, there is lots of data that we can take advantage of. So there's a data explosion, but there's almost if we can't organize it, if we can't present it, if we can't get value out of it, it's just data and it overloads our underwriting. And so we have this data explosion which should make us be very, very good underwriters, but we haven't been able to take advantage of. At the same time, the risks are getting more complex, right? Every small business now has some sort of cyber exposure, considering all the clouds and stuff of their own. The same is true for any middle and large business things that used to only be a commercial enterprise. Now I have a retail component through various E sales and whatever else, and then you have all the changes with global risk and weather events and along those lines that are just accelerating. So against that backdrop, we know that we need to improve efficiency and effectiveness tremendously so we not only have to be able to process our underwriting faster, but we got to get to better decisions. And the only way that we're going to do that is really start to look at these technologies, to really look at the data, to really look at the analytics. And where it all comes together is the artificial intelligence to help us at the different steps of the process as we kind of think about and consider it a little too far there. So. One of the key areas that we see is that we do surveys of underwriters from time to time and look at where they're spending their time and how they're spending the time. We looked 13 years ago. And what we found is the underwriters were spending a lot of their time on non-core activity. So what are non-core activities? Non-core activities are spending time gathering the data. Organizing the data, entering the data into their system, getting the data ready for them to actually go do the analysis. We just we just surveyed the underwriters again. And what we see is they're still spending time doing those same things. They're still spending too much time on data gathering, data preparation and data entry, as opposed to actually analyzing the risk, evaluating the risk and putting together packages that will drive profitability and solutions for both the insured and the carriers. We need to get them doing what we want them to be doing. And the way to do that is through these data and through these analytics. So what we look at is when we take a step back, the first thing to recognize is the reality is insurance really hasn't changed in some 300 years. So 300 years ago, when it was time for a submission, the broker or agent would pull all the information together. They would walk it down to the carrier and they would have this big file folder in them and the underwriter would go through it. You know, Then we started to mail it. Then we sent it through Federal Express. Then we sent it through fax. And today we take those same paper documents and we make digital versions of it, Right? We don't turn them into data. We turn them into PDFs. And rather than sending them by fax, we send them by email. And then what happens when they get there? What the underwriter does with those PDFs is what they used to do. Well, I remember I've been around here long enough. I remember the time when the height of underwriting technology was like the seven part file folder. Right. And so the long run went into the first part of the application, went into the set, went into this part, and then the financials went into Section three, right? And you had whole rooms or even whole floor sometimes dedicated to file cabinets that would hold these. So what have we evolved in 300 years? Well, now when those PDFs come in, what do we do? We have electronic file folders that are set up with the same sections of Oh, the Lost friends go here, the financials go here, the applications go here, correspondence with the broker goes here. So we put these paper documents into and what do we do with those paper documents? We pass them along the process and those paper documents, PDF's. But they get opened up again and again all the way through the underwriting process. And so. We haven't really transformed underwriting. We've pretended to go digital and move that paper into electronic form, but we've never really made the underwriting process digital. And so what we're seeing with the technologies now is now there is a chance to actually rethink that process completely because the technology has gotten to the point where we can use it in different parts to, first of all, help us to get that data right from the beginning, improving our interactions with the customer, improving the accuracy of the data, and digitizing that data right from the beginning so that we can use it. We're finding ways to apply that AI in such a way that it can organize and repair data. But in a human responsible way, we want to make sure that it's not making decisions that would the regulators would frown on and whatever else. So we want to make sure it's happening in a responsible way. And what we're seeing is the AI is now mature enough that we get real returns actually incredibly fast. And so. This rethinking of the process from paper based to digital is the big transformation that we see, and we see us using AI along those pieces to do. The transformation to be responsible and interactive is how we work with humans and to really drive real results in real. So how do we see this coming together? Well, the first thing we've seen is there's been a lot of experiments today. We have lots of little pieces of where we recognize the value of the data and analytics and various carriers have tried to do solutions. Right. So that one carrier and they have like 92 I'm serious about this. They had 92 different models their middle market underwriters were supposed to use in the given situation to help them make decisions. But each of those models were built in their own way. Each of those models were its own little platform and whatever else each of those models was. Something was supposed to go and it wasn't usable, it wasn't viable. It hadn't been designed for a human experience. And. We've been saying this for a while where there were these different plays and different parts going on where we're trying to evaluate different tiny pieces of a solution. And what we recognize is what's really missing is we need to take a step back in order to move from paper to digital. We actually need a third generation of capabilities that we need to provide consistently and holistically. So what do I mean by that? Well, if you look at the life cycle of technology and underwriting. The first generation that we focus on. And this goes back a ways into the eighties and nineties, early 2000 was getting the platform systems right. Now originally we did it with Big iron and now we're having much more than we went through server systems and now we have much more stronger platforms. But it was the core policy platform, the platforms that hold the data necessary to control the policy so that we can manage an efficient. And that made sense. That was the first area in which we needed to apply technology. The second area in which we need to apply technology then was how do we handle the management of work? So how do we pass work around? And that was the second generation and we saw that come in in the 2000 and the teens and whatever else was all workflow systems. Right to workflow systems that allow us to that pass those digital files from person to person so that we could work them in a cohesive way, manage our work all the way through and get it. And most carriers, this is where we are today. Now we have we may have a mix of legacy systems and other systems, but we have a mix of platforms and essentially desktops or workflow systems that manage our work. But if we look at these systems, the only manage like an average submission has anywhere between 305 hundred data fields, especially when you're talking about commercial insurance. These systems manage about 50 fields of that data. And that's the data we put in and that's the data we use to rate, and that's the data we use to match. What we leave behind, though, is all of that other data and information. This is what we've tried with analytics or point solutions or pulling in from third party, trying to get it into our workflow systems and trying to get into our policies. But they're not set up to do that. They're not set up to manage all that extra data. And so the leap that we're seeing now that is getting us away from experimentation to transformation is the fact that we now need to add a third generation of the next leap, which is where we bring the data together. So a way to hold all of that extra data and the third party data on our internal data, whatever else, all that other data that does not belong in our policy or our workflow systems, but is the data that the underwriters need to actually evaluate. And then using that data to help automate our processing with either robotics integrations or AI using it to enable the underwriter make decisions by having models and things along those lines, or being able to automate all elements of costs. And this is this kind of third generation of digital underwriting that we see as the leap that needs to happen for us to take advantage of the AI. So it's a different type of thinking to get us beyond that paper base and truly think and truly evaluate things as a digital underwriting process rather than a paper driven process. And that's the big leap that we see starting to occur. And we see this being implemented, right? This is not future. This is coming to pass in very real ways. A number of different carriers approaching slightly different ways where we're going to do and it's one of the exciting things of our partnership with AWS, of course, is going to talk to about a second is we built one of these platforms that enables us to help carriers very, very quickly, kind of get to this next step by introducing this third generation platform that needs to work in conjunction with the investments that have already been made at the first and second generation. And all of this as we start to move into this digital world, Obviously, we still need to consider responsible AI through all of this because there is the possibility to use AI for elements of data extraction, for elements of data automation, and for elements of, if not the actual decisioning, preparing the information for decision. As we do that, we want to make sure that we're doing it in a safe and controlled way that is treating all of the submissions that come in ethically and responsibly. And so there are new responsibilities with this digital platform as far as thinking through the testing, thinking through the preparation, testing to make sure that we're moving in the right direction. Sorry. And the big thing about this, guys, is that. When we think in this way, when we think of moving from a paper based environment, which is where we still are, to a truly digital environment and we take advantage of these like this. Third, we treat it as a platform, not as a point solution, but as a platform that we can utilize in multiple places. The costs of these technologies are much lower than what we've had to do with our first generation and second generation stuff. Big giant policy platform replacements, big giant workflow really are costly, timely and take a long time to realize that. What we're finding is that the cost of AI and the cost of these data technologies have fallen dramatically. And so when we apply them in the right way, the investments are less than what we see in the other elements of the platform. And the speed is tremendously faster in which we can start to get value, especially when we think about it not as an individual solution or not as a point solution, but start to think about it as reusable components so that we have a data layer that is reusable, that we have analytics that go against the data layer and a consistent way and drive these elements. So that platform approach really gets us a lot of value and start to drive down costs. And with that, actually I'll stack up a slide with that core. I want to turn it over to you because you've been driving Accenture built out of this platform with AWS now, has us ready to help carriers transform into this new truly digital underwriting world.
Corey: [00:26:01] Thanks. Thanks, Michael. And thanks, everyone, for taking the time today to learn a little bit more about how we're progressing. AI and Underwriting in insurance. Just a quick intro. So my name is Corey Barker. I'm managing director of Accenture insurance practice. I'm the product owner for our asset initiatives to build insurance specific use cases, leveraging the power of native cloud capabilities. I'm originally from Australia. I moved to Chicago about eight years ago and we've been driving digital transformation at our PNC insurance clients this whole time. So let's talk about how we bring this to life a little bit. So our vision when we first started this discussion with AWS, our vision was an enterprise grade processing platform for unstructured content that will decouple intelligence from the core insurance processing system. What we've done is we've created an end to end experience. But given that good underwriting is part of an insurer's DNA, we understand that some insurers have already invested heavily in certain areas. So what we did was we built an enterprise platform that is modular and allows insurers to bolt on existing capabilities or where there is use of other point solutions being able to leverage those also. So let me just take a few minutes to talk you through these individual modules. A Mindful that we won't have a ton of time to deep dive on the content here in the solution, but I think Paul has mentioned before we will share the content, our contact details and more than happy to to set up some more time to deep dive with any of you on on the on the capabilities that exists here. But I do want to give a little bit of an overview of the solution and really thinking and then talk through a little bit about next steps on how you can bring it to life, whether it's with this or even just thinking through how to think about infusing AI into your underwriting processes. So if we look at the end to end lifecycle here, we've got a basic, I'll say a standard capability across submission all the way through to to quote as we think about the lifecycle of of a submission, but we'll take it in different parts of the sub process and I'll talk about the process and more importantly how the platform and capability can help insurers carry insurance carriers with their day to day jobs. So if we look first at the submission and extraction process, although there have been some advancements made in broker and agent portals and there has been some efforts made in standardization of submissions, as Michael mentioned before, underwriters are still receiving initial submissions from brokers either directly or they have to manually retrieve it from a system. Then there's a lot of still manual, quite manual processes that are required to continue to work on these processes. So what we've done is we've designed an ingestion engine that meets the brokers and agents at their point of need, giving them the ability to email content such as submission documents and lost runs and intelligent email and intelligent ingestion modules can route those emails correctly to the right people that can ingest the documents automatically. So there is no need for manual rekeying of submission documents. And when we talk about intelligent ingestion, this isn't just OCR. We've built this out specifically for commercial lines, insurance products and and we're seeing 95% accuracy of hundreds of insurance specific documents and insurance specific context, which is really important to bring as we bring in that data. So the ability to be able to see what's policy information versus risk information versus if we've got our an aerospace policy versus property policy, really being able to understand the context of what's being submitted versus just kind of data data on a page. So once we have that submission data, how do we how do we look to enrich that? So as Michael mentioned before, underwriting work environments have become more and more complex as the additional systems that are being added to. There's new tools that are being added to. There's new additional data sources that are being required to to for underwriters to make these informed decisions. And it's really overwhelming when you think about how much how much data is accessible, but how much data needs to be reviewed to actually come up with a a quality decision. And right now that's being put the owner is being put on on the underwriters to do all of that. And we're seeing upwards of 30 to 40% of an underwriters time being being taken just in gathering that data and not using the time to actually make informed decisions. So what we've done partnering with the team is we've built the ability to infuse third party data providers who are focused on helping insurance carriers. We've already built integrations with some traditional insurance data providers like Dun and Bradstreet, like LexisNexis. Precisely. But we're also leveraging Accenture's broader data ecosystem of over 950 data providers, which is somewhere around 20 terabytes of data, plus working with the AWS, the AWS team and that data exchange. So that allows carriers the flexibility of infusing that third party data, both from an underwriting decision perspective, but also from a research perspective and understanding what truly is, is is driving underwriting or at the end of the day, ultimately how can you make better decisions? And that takes us to the assessment process. So what we're seeing is, although there is a desire by insurance companies to allow for further straight through processing, the nuance with commercial underwriting is there is still somewhat of a need for underwriting expertise. There is the underwriter is still at the heart of the decision making, which makes complete sense because of the amount of experience underwriters have in the industry. There's certain things that you still just, quite frankly, can't glean from other than underwriting experience. But what we're trying to do here is not necessarily replace that underwriting decision, but help underwriters make those decisions. And what we do and how we do that is actually giving contextual feedback and information and insights that underwriters can then use to help make the decision easier. So a couple of examples that we've got here. One is comparative analytics. So how does the company that's being reviewed compare with its peers as you look at companies that are in the same industry? What do they look like in comparison to size of revenue, size of employees, location, risk factors that are applicable to other other companies in that space, Really just a consolidation of some of that data to help with decision making. We've also helped with appetite triage. So we've we're assisting underwriters to with a score, a submission score based on carriers unique underwriting appetites. So we understand that each carrier is different. And what the focus is from an appetite perspective may differ from Carrier to carry out. What we've done is we've actually built a low code rule capability that allows for easy upload of carriers unique needs, which then helps underwriters focus on the submissions that fit better with within a carriers appetite. Lastly on that, we we've also helped with a propensity model so we understand some independent agents and brokers will send submissions in quotes to carriers without potentially without the intent of buying from that carrier. Sometimes there's some shopping around, sometimes there's comparison of different quotes. So the third the third example that we have here is a score based on propensity to buy, and we use things like history of engagement of that particular broker. We look at what the brokers are focused on in terms of the business that they're bringing to the carrier and ultimately coming back with a score from a propensity to buy score, which again helps with the triage of which submissions should I be focused on first. And it's the main driver of that is if there's a broker that you know, or an agent that perhaps is is not you don't quite have the same relationship or maybe it's not as strategic, then the score will be lower than perhaps an agent or broker that you do a lot of business with. So those are just three model examples of insights that we're giving back to underwriters that aren't necessarily saying whether something should or should shouldn't be underwritten, but just helping with that, with that triage process, that underwriting process to then help them make decisions faster. Finally, what we're doing is we're packaging all of this capability. So the ingestion capability, the submission capability, the the third party data ingestion capability, these scores and insights that I was just talking about, and we're packaging all that up into a visualization and portal. So one pane of glass for underwriters to see all of these modules, there's workflow capabilities built in for referrals. There's the ability to use that as a kind of a source of truth really, for decisioning. And what we're seeing is that that pane of glass is really helping with the speed in terms of decision making and really just being able to help as that one stop shop for underwriters to view the data of their submission. Finally, we've we've spent time working with the traditional core vendors within insurance policy administration systems like Guidewire and Duck Creek to automatically take the decisioning from from these outputs at taking those inputs into the platform with integrations to create quotes in those core systems of record that will further speed up the submission and quote process. And what that does again, is continuing that that mantra of automation, reducing manual rekeying and an end to end platform that can help with decisioning. It's just another example of how we're bringing that speed to get response times back to the brokers and agents that the carriers are working with. So as I've mentioned upfront, this has been developed as an end to end solution. But you can see how several of these modules could be deployed individually. For example, we have some carriers that are focused just on the automation of ingestion. That's a bottleneck that they've found that they need help with in terms of just focusing on that particular area. There are others that have already done a lot of investment in in the modeling capabilities and third party data and they just need the the ecosystem enabler, that portal, that single pane of glass. And we're working with them to integrate their third party data sources and models that they've already built into that portal and helping with from a workflow perspective. So there are a few different ways to think about this end to end process and understanding where your particular point of need is and starting there and kind of working out, focusing on that and kind of expanding on it from there. But what we've found is end to end, we're seeing up to a 50% reduction in cycle time. As you can see, if you look at all of these individual pieces and you add that all together, it's actually really powerful from a speed, a speed perspective of response and then ultimately with a speed of response and increasing quality. In terms of data retention, we're seeing improved broker and agent experience, which obviously helps with retention. And we don't have the data on this just yet. But anecdotally we believe what we're driving between 2 to 6% on loss ratio reduction purely because of the ability to bring in the additional third party data, having these standardized processes really helping the underwriter move from data gatherer to decision maker is helping insurers drive down loss ratios. So this is as I assume you're probably always thinking, this is great. We don't have a ton of time to think through this end to end transformation. You know, we've seen transformations like this take years to roll out, potentially millions of dollars. And so I just want to make everyone think through how to think about this differently. And I think to Michael's point before, you know, the nuance today versus perhaps 3 to 5 years ago is these are all cloud enabled cloud native technologies. And what that means is we can be we can be extremely focused in a particular area and we can be extremely fast in terms of how we develop and go to market. And so as we think about delivering value and accelerating that value, you know, here's an example, and this isn't prescriptive, but it's just an example of a way to think through how to deliver value quickly. So we've got the 30, 60, 91, 20 kind of framework here. I'll just quickly talk through an example of how we've how we've worked with our carrier on on how to think through this. So within the first 30 days, we can have a proof of concept developed. If we were to take a particular sub process or use case part of the end to end flow that I talked about before, if we took ingestion, for example, we could actually stand up a an ingestion capability within 30 days, start to get some of your documents through and show that it can work because a lot of a lot of the times now what we're seeing in terms of the starting point of change isn't necessarily that the technology isn't there. It's the inertia of decision making sometimes. And a 30 day proof of concept usually cuts through that. There's a there's an ability to be able to prove that that this can work. There's an ability to be able to show stakeholders within your companies that that this is something that can add value. And it's more than just theoretical tube processes. It's it's actually showing this is how the solution would work. And what we can actually do alongside that is think through, okay, while the POC is going on, on the assumption that it's successful, what does this mean in terms of target state? And really thinking through what what the underwriter of the future is going to look like. And to Michael's point, before just kind of confirming that that vision on the end state as we look forward over the subsequent kind of mini milestones we can over on day by day 60, we can have the proof of concept validated. We could have an initial business case on what that would mean to your current state and and moving to a future state, what that would mean from a business case perspective on those drivers that I talked about before. So, you know, speed and turnaround time retention from a broker and agent perspective and ultimately better underwriting quality and decisioning, those all form part of a business case that we can help you work through on defining how how this is going to be enabled from a cost perspective. And then as we look at 90 days, 120 days out, that's when we can really start to get really focused on your specific need. And so being able to say right, bye bye after three months, we can have a pilot defined and after four months actually have something working in production. It's actually really, really amazing to see that this can happen and be the the mindset shift that it allows insurance carriers to think through on, okay, well what can I do next and how can I do something in the next 30 days? And what else do I want to try and prove out to continue that that more efficient underwriting and really moving into an agile delivery model and an Agile mindset. So this is just an example of how to how to accelerate that value, how to think about starting to transform the organization, the underwriting organization incrementally instead of like a large scale transformation program. So I touched on a lot. I know that there's I'm sure there's a lot of questions. I would be more than happy to do a demo of the solution. Deep dive in certain capabilities. Definitely feel free to reach out to myself, Michael or Aarti, but I'll throw it back to you to host the questions.
Paul: [00:44:14] Yeah. Thank you, Corey. Thank you to Michael and to Aarti as well. Great to have you all back with us. And we will we will dive into some questions. Yeah. So people have been taking advantage of the Q&A box and obviously we still have about just close to 15 minutes left of the webinar today. So if you do want to pose a question, please use the Q&A box on the right hand side of your screen. Just type your question in there and we will attempt to put it to the panelists at time permitting, of course. So I'm going to start with one. I'm not going to read out the names, by the way. So if anybody wants to stay anonymous with their questions, don't worry about that. I'll just quote the questions. So one that came in a little bit earlier was what are the accuracy rates we have witnessed related to intelligent ingestion solutions? So I don't know if any of you can can pick that one up.
Michael: [00:45:08] Corey, you want me to take it?
Corey: [00:45:10] Yeah. You go Michael.
Michael: [00:45:13] So we're actually seeing and what we've seen is a change in technology. So the old technology is used to see the accuracy is kind of hovering in the seventies. You might get a bid to push them in the eighties. The new technologies we're seeing is, first of all, it's extraction ability is far higher and we're getting well into the nineties in both its ability to find the data. So what it's attempting and then the actual accuracy of the results that it's getting out of is well into the nineties across a wide array of different document types that we've seen so far.
Paul: [00:45:48] Okay, Excellent. I'm going to take you to another one that we saved from earlier. So does the solution. And the example given is a AW portal provide a view of the account or the portfolio level risk exposure. Who can take that one?
Corey: [00:46:08] I'll tag team here Michael so yeah, Paul. We focused there's two focuses. One is on a particular submission. So being able to bring all of the data and the capabilities to, to a particular company or, or client that's part of that submission. But also we also have the ability to see at an account level or a portfolio view for a particular underwriter. You know, what is part of the overall account portfolio. So where, where are we seeing condensed areas of risk? We overly we're looking at particular whether it's regional, geographical or specific industry focus. And so it gives the underwriters the ability to to look at manage underwriting at a portfolio level as well as a submission by submission basis.
Michael: [00:47:09] Yeah, and I'll take it a little bit further. Right. So this is part of the transformative nature of moving from paper or digital paper to truly moving to digital underwriting. So think about what an underwriter does today. When a renewal comes through, what's the main thing the underwriter cares about? Is this risk different than it was before? Right. So how do they do it? Well, most underwriters, they'll have two screens going and they'll have two PDFs open and they'll be reading through the two PDFs. And meanwhile, paper notes, they'll be trying to track what's different across the two to figure out how they need to adjust price or a new submission come in that they saw it last year. They're trying to look at what it was like last year. When the data is digital, the experience can be completely different because we can put that data side to side, we can flag. What is different of this year versus last year? We can flag. If I pull up another 100 companies that are like this one. Where does this one fit in, in this peer group? Where is it riskier? Where is it less riskier? And feed that insight to the underwriter as you really start to think of if the data is digital and that's the internal data, that's the data that you got from the carrier and the broker, it's the data that you get from third party. If that data is digital and how I can arrange and organize that for the underwriter, you completely change the experience from what they're having to do today. And that's what the platform has. The platform has examples of all of this, right, of comparative analytics of year over year, analytics of bringing third party pieces together. And so it's all part of this re-imagining how underwriting can actually function when you get out of the digital paper processes that we have today, which is what makes it so exciting and what's going to make it truly transformative for the carriers that take advantage and really give them a first mover, a first mover advantage. Because this is not just about efficiency. This is not just about getting the quotes out faster. This will also fundamentally help the underwriters to make better and more consistent decisions. And that's that's why it is so transformative and powerful.
Paul: [00:49:37] All right. Excellent. Thank you. Corey, I'm Michael. One more question that I'd saved from earlier during the conversation, and I see that some more questions are filing in. Please do keep them coming up around 9 minutes left. So get your questions in through the Q&A box now. But yes, so one more question that I saved from a little bit earlier, which is what is needed from the carrier's end to get the posse going. Can someone elaborate?
Michael: [00:50:06] Corey, you want to go. You want me to?
Corey: [00:50:07] Yeah. So I mentioned that a little bit before, Paul, but the based on the module or the specific sub process that we'd want to tackle first, if it's the ingestion process, for example, we would look at some example submission documents, even just what's been what the carriers have received over the last week, we can use that. Or if there's a particular line of business, obviously we can tailor the pure data to a particular line of business or a particular geography. So just having those documents to start with provided from the carrier, we can we can then show very quickly how, how those can be ingested, ingested and what that would look like in the portal. We have also worked with carriers in the past where they're apprehensive to initially give example submissions. So we've we've even used just agreed on test data or provided from the carrier to prove that the capability works. Michael, I don't know if there's anything else you wanted to add.
Michael: [00:51:27] Yeah, I think there's a couple of things I want to add. The first part is because it takes a platform approach, right? So that can hold its own data, has analytics, workflow and those sort of pieces. And it's designed to work with existing workflow systems and existing policy systems and stuff that you have, right? It's like you can figure out where those integration points are that you want to do. You can actually go. Corey is absolutely right. You can go very light and you can go very fast, right? As far as the different elements that came in and whatever else, I mean, you stood up ingestion pilots in very short windows, right, of taking the data in and whatever else. You can figure out how the data and stuff can go in very short windows because. You don't have to rebuild your whole infrastructure. And the other part is because you can do data, because you kind of have that big data kind of sitting in the middle. You can sit over top of messy environments. You do not have to have one workflow system or one policy system. We're working with a carrier now. We've got a whole bunch of different policy systems and a whole bunch of different workflow systems that we're having to sit and integrate with. But because we can figure out the data flow pretty clearly, we can use a mix of APIs and bots and other integrations to connect into the legacy challenges that are there. Like this can move independently of all that stuff. And so and then the other piece is it was built with a lot of insurance knowledge. So sometimes a lot of other platforms are a lot of technology. They have great technology, but then they have to come ask you what everything means. They don't know all the unique elements of insurance and whatever else. This was built with deep insurance knowledge to start. And so we don't have a lot of those businesses to the same degree and stuff as well, because a lot of things are already there and a lot of things are already thought through with people who understand the industry quite well. And so the ask is much lower than you would see with most traditional, most traditional new technology implementations.
Paul: [00:53:40] Okay. Now, we do have quite a lot of questions that have come in. I'm very conscious of time. We've only got 5 minutes remaining, so if you could be conscious of that with the answers that you're giving. So the next one that's come in is please shed some light on how insurers can leverage AI and analytics in a greenfield environment. Aarti, I think this might be one for you.
Aarti: [00:54:03] Yes. Thanks, Paul. Yeah, I'm happy to answer that. So I think from an AWS standpoint in particular, right. So I think the beauty of AWS services is meet the organization, meet the insurance company on where they are. We always work backwards from the customer. So specific to AI ML, we really see three flavors that I would like to kind of point out when it comes to services and insurance. Right? So the first one is the managed or finished services that are available within an AWS environment that you can write your APIs to in order to point to those services to the data that you have in the data link right. And just a quick example of that would be like Amazon comprehend pointed towards customer communication or textract to extract the data out of these underwriting documents that we were discussing. The second flavor would be kind of build your own or bring your own where your data scientists can use models like Sage Maker to kind of either kind of bring your own models or you could even quickly train and deploy new models into machine learning and just kind of build as you go. So the sage maker capability is kind of no prior experience needed type of thing. And then the third one, which I think is more relevant to the discussion today, are the partner developed AI models, right? So the one that we're discussing today in partnership with Accenture is a great example of that. We have several other partners, but just to kind of stay focused on the underwriting solution in particular, that would be the third flavor where you can just kind of quickly bring a partner in through a ready made solution and also use their deployment. And some of the things that Corey and Michael were pointing to kind of use that and then run with it. So it really depends where your organization is. But like I said, kind of meet you where you need to and what best suits your needs.
Paul: [00:56:03] Okay, let's see if we've got time for one or two more. So next on the list is how can brokers be supported in this transformation? So broker side, who can pick that one up?
Michael: [00:56:17] So two of the same technologies that we're talking about applying to the carrier actually have similar life in the brokers. And we actually are seeing some examples where we're doing that, especially with some of the data ingestion and some of the analysis processes obviously are different for the brokers, but they have similar types of problems and challenges. And I saw the other part just jumping down as the question was if we implement a platform like this, does it create change for the broker? And the advantage is that it doesn't. It's the technology is designed to work. However, the broker wants to work with you. So just to jump ahead to that one too.
Corey: [00:56:56] Yeah, I think to add maybe just one one thing to add to that, Michael. The intent of this is really to make the brokers experience even easier. So for them, the main difference would actually be not in the way that they interact today, but actually their expectations on speed and response. So because we're automating, we're speeding up a lot of these processes, We're getting to decisions faster. The broker experience will be will be for the better. But in terms of what they're expected to do, there's really no change in terms of how they would have typically interacted.
Paul: [00:57:41] All right. Thank you. And yeah, I don't think we have time for one more, unfortunately. So just a quick one from me to say thank you to everybody for joining the discussion today. And thank you, of course, to Aarti, Michael and Corey for a tremendous talk. We hope that we've addressed a lot of your questions and remember, you can always reach out afterwards as well to to find out more. I've been Paul Lucas and on behalf of Insurance Business and Accenture. Stay safe, be kind, and we'll see you next time. Have a great day, everybody.
Corey: [00:58:11] Thanks, Paul. Thanks, everyone.
Michael: [00:58:13] Thank you. Bye.