Thursday, May 23, 2019
10:00 am EDT | 03:00 pm GMT | 04:00 pm CET
The AI Pathway to Transformation in Commercial Insurance
Thursday, May 23, 2019
Squirro and Accenture discuss how Commercial Insurance can be transformed leveraging AI
The most disruptive changes in the insurance industry will be led by the advances in and utilization of key technologies. Primarily, it will include the adaptation of AI to create desirable products, harness insights from new data sources, optimize processes and costs, and improve customer experiences. According to a recent Accenture report:
- Digital transformation grants insurance companies exceptional capabilities. But it also creates enormous expectations.
- Amid these rising expectations, insurers are investing in digital technologies, raising the question of how leaders will set themselves apart.
- Insurers looking to differentiate themselves must be aware of technology trends that will characterize the “post-digital” future.
This same report says that insurance workforces are becoming Human+: each worker is empowered by their skills and knowledge plus a new set of tech-driven capabilities. Now, insurers must adapt technology strategies to support a new way of working in the post-digital age.
Watch this webinar with Squirro and Accenture to learn how your teams can be empowered with AI-driven insights and recommendations, creating an environment of Augmented Intelligence.
ERIC ROBUSTELLI: Hello everyone, and welcome to Squirro and the webinar, The AI Pathway to Transformation in Commercial Insurance. My name Eric Robustelli and I’m the North American Head of Sales here at Squirro. We’re joined today by our Chief Sales Officer, Miguel Rodriguez, and Accenture’s very own Silke Genuit and Sarah Steiger. Please, if you do have questions during this webinar, submit them into the questions panel on the screen and we will answer them at the end of the webinar. So, thank you very much all, and we’re looking forward to this very exciting webinar today. Without further ado, Miguel, please take it away.
MIGUEL RODRIGUEZ: Thank you very much, Eric, and thank you, everybody, for joining this webinar. Today we will be talking about innovation and insurance. Some of you might know already, this picture here, everybody knows this, sometimes we are too focused on our day-to-day work, and innovation is not always the first priority that we have. But we hope that with the content of today’s webinar and the insights of use cases that we can deliver within insurance, we are able to bring this to the top of your mind and work with you in bringing innovation to the commercial insurance space.
First of all, I would like to start with an introduction to The Evolution of Data Analytics and AI, basically from Excel to AI. How has data analytics evolved? In the 1990s a great analytics tool called Excel was introduced to the world. Excel is a great tool to analyze numbers, but it is also a tool which requires very specific skills. It’s also prone to human error and not easy to use, plus it always shows you just what happened. That’s why in the 2000s companies such as QlikView and other BI providers, business intelligence providers, they started to provide a visualization layer for structured data where you had the possibility, through dynamic dashboards, to understand, closer to real-time, what was happening. Also, users with fewer skills in Excel had the possibility to analyze numbers. From 2017 onwards we have seen a shift towards people not just wanting to analyze the data and understand what happened and what’s happening, but also with the predictive element in mind, to understand what will happen. And this is why today we are talking about actionable insights and actionable recommendations for users not only based on structured data but also a large amount of unstructured data that is available in the companies.
If we look at the evolution of IT, we see a similar evolution here. In the 1990s data warehousing was the big topic, it was all about storing data from multiple databases. Then we had the hype around Big Data and Big Data analytics in the 2000s. Also, in the last couple of years, where the main topic was to handle large amounts of data from various data sources. Similar to before also in the last couple years, we have seen a new trend coming up and it’s the topic of Artificial Intelligence, or also machine learning, wherein the IT landscape they look for solutions to have the ability to learn, and that acts like a human and provides actionable recommendations for people so they can be leveraged in a lot of easy to understand ways.
If we look at our day to day life, AI and machine learning have already found its place. Why is it adopted in the retail space already better than in the B2B space? There is a very simple reason for it because it’s not intrusive. The usage of AI and machine learning is easy and comes in simple to understand ways. So, you receive notifications on your mobile phone when your next bus is coming. You receive information in context of what is interesting for you on your mobile phone as well. On a daily basis with prefilled good news, customized to where everything that is relevant for you, you have Google assistants and the Alexis of this world. These have seen amazing growth in today’s living rooms because you can just talk to them and they provide you information. And all of this is based on AI, as well. We see self-driving cars, we see headphones that can translate in multiple languages in real-time, and watches that leverage information to monitor your house, to provide you recommendations on what you need to do. This is possible, specifically in this space, because there is a lot of data available and this data has been used by these companies to make machine learning and AI available to everybody in a way where you don’t need to have any specific skill because the information is just simply provided to you in many ways that are very natural for a human.
We see that in the business space this will become very, very relevant and it’s not in a couple of years, but it’s already tomorrow because next year is already 2020. As we see here in this statement from Forrester, “Companies that are truly insights-driven businesses will steal $1.2 trillion per annum from their less informed peers by 2020.” So, we are convinced that it’s really time that companies embrace today’s technology and – in the same way as the technology was embraced in the retail space – in the B2B space, businesses think about how they can manage this technology to empower their workforce. People like relationship managers. How can we use AI as an example to provide to these people seamless recommendations so they can also work with this highly sophisticated technology?
This brings us to Squirro and the problem that we tried to solve, which is really in this context. In the past, we saw that knowledge workers, people working in relationship management, customer service, and other different areas, they were working and spending a lot of time finding the right information. Information was not coming to them as we see it today on the devices that I explained. This was a very time-consuming task because data, and specifically the unstructured data, is siloed and scattered across the organization. So, in the same way, as we see information coming to us on a mobile phone, where it’s put in our context and it’s been analyzed and the context is understood, we should see the same in the B2B space, and this is what we provide. We bundle up this information, we gather this data, we analyze it we visualize it. Then we don’t just deliver a complex solution for the end user but integrate with third-party applications so that users that don’t have specific skills around machine learning and AI can leverage this information too, in a seamless way.
Squirro as a company has positioned itself in the Augmented Intelligence space, really with a mission to augment and not replace, to provide information in a tangible, actionable way and in a real business context, as I was explaining, and to be very pragmatic about the topic of Artificial Intelligence. We have been doing this all over the world with customers and with the support of partners like Accenture that are today helping us here in this webinar. With the backing of great investors such as Salesforce Ventures, ICOS Capital, Finch Capital, and the Hammer Team. Our core offering, as I’ve outlined already before, is really around these three topics of gathering data, analyzing data, and providing an actionable recommendation within core systems. And to make this more simple for everybody and not start from scratch, we have built AI, Augmented Intelligence applications which solve real business problems and have predefined connectors, predefined capabilities, and predefined visualizations so that users can start from day one to take advantage of both machine learning and Artificial Intelligence. During this webinar, we will learn more about this and I will show one specific example.
So, I already mentioned Accenture, it’s a great partner of ours, a gold partner status. We work together because we think that we have a strong joint offering that we want to bring to the market. So, on one side, I already mentioned what Squirro is bringing. We have this powerful Augmented Intelligence platform that comes in a variety of pre-built configurations, we offer implementation support for our partners, make sure that the platform continues to evolve, is continuously fed with new updates, and the maintenance is guaranteed. Accenture doesn’t need a lot of introduction, a well-known company all over the world, brings really global innovation architecture to support design, thinking, and rapid prototyping. They have access to Squirro certified resources, have been onboarding a lot of people within the company to be able to deliver and support the implementations that we do together, and they obviously also have skills to extend and expand the platform with new capabilities, and also bring the very important vertical specific knowledge into the application to build more predefined applications that can be brought to market and have customers take advantage of it. So, combined, we have a joint project delivery team that can be onsite to support the project implementation. We have full support from integration and development, through data management to rollout and thereafter, rollout service and maintenance, and we can really help you, together with Accenture, to develop an AI strategy and make a blueprint of the use cases that are relevant for you. We can help you get there with predefined architecture, a roadmap that we can develop together, and applications that we are happy to develop together and define together with you during our workshops and the joint meetings that we will have. So, with this, I would like to hand over to own Silke, as she will guide us through The Evolution of Insurance.
SILKE GENUIT: Thank you, Miguel, I’m happy to. So, when we talk about insurance, we would like to recap a bit on the insurance business model as such and start with a somewhat bold statement, saying that the insurance model hasn’t changed much in 200 years. And it’s basically capital placed against risk as tailored products, using the information to sell to customers through channels and generate returns for a company.
Since this chapter is called ‘evolution’ you would think that now actually everything is changing. So, capital is changing. We have much more alternative capital, we have self-financed capital, we have peer-to-peer. Risk, as such, is changing a lot as well; we have micro risk, cyber, pandemic, even paper use, et cetera. And tailored products are changing, so we have episodic, hyper-tailored, parametric products. We have information that is changing a lot, as well, instrumentation, big data, social. We have customers who are changing, their attitudes and expectations are changing. It’s much more multidimensional than it used to be before. The channels are changing, it’s much more digital, as Miguel mentioned we have online, we are embedded, there are more agencies, and the returns are changing, not exactly for the better, so it’s shrinking returns. And for the company, there are more and more companies out there, insurtechs, lots more incumbents, and much more disruptors out there.
Looking at the consumers specifically, as our research recently stated, consumer attitudes regarding insurance are also changing. They are influenced by the GAFA, so, the Google, Amazon, Facebook, the Apples of the world, they expect a similar experience when interacting with their insurance that they’re used to from interacting with Google or with Apple, or with Facebook, all the digital channels they’re used to. So, they would also consider, and they have trust in them, 38% of the surveyed said they’d consider buying insurance from them; 57% would also be willing to share more information in return for added benefits, for example, on the health activity status in return for lower premiums; 74% would be willing to entertain computer-generated advice in regards to what kind of insurance they should purchase.
So, GAFAs and insurtechs are really becoming much more central to people’s interactions in people’s lives. So, insurance, as a business model, would need to react to that. And we see this happening a lot already in the US, in Asia, for example with Trov, there is much more on-demand insurance that you can turn off and turn on when you need it. And even looking into China with ZhongAn, that’s already pay per use insurance in their online shopping across sector partnerships. So, there is a lot of already happening and that is not really happening with traditional insurance, as these are all more or less new incumbents that capture part of the value chain, and they capture essentially the connection with the customers via these channels.
Also, risk, of course, is evolving and there are much more new challenges, but also, of course, opportunities out there, especially for insurers. So there are many new exposures with driverless cars; autonomous driving is going to be a whole new question regarding liability, cybersecurity, sharing economy. Everything related to digital ID and assets is going to be interesting to look at and see what new products come out of that covering these new risks. There are new ways of funding, cashbacks, P2P insurance, and new ways to mitigate the risks. Since we’re used to Spotify, we’re used to subscriptions, and maybe we don’t want to pay for everything in advance, but we want to subscribe to some sort of insurance coverage, et cetera. So, it’s different ways to mitigate risk, it can be much more automated, much more invisible, much more personalized and much more situated to my actual current situation where I’m in, so real-time protection.
And the size of the prize as our research stated, there was $1.2 trillion mentioned by Forrester regarding insurance, especially commercial insurance, that’s our focus today. We believe there is a prize of $69 billion in commercial insurance who get it right, to capture the market share in this line of business. So, there is a significant amount of revenue potential in there.
We believe these opportunities spread around five revenue growth areas, specifically. First, in regards to the increased market penetration. Segments that have been very difficult to reach profitably have now been much more available and reachable via the new channels and technologies. There is much more new risk out there and hopefully more new products that will be able to cover these risks; there is a prize there, as well. And then new partnerships and relationships with the traditional insurers to cover the reach and the data insights of the GAFAs, for example, to connect and basically directly, subsequently offer these new offerings. Then there is monetization of the “new.” For example, where the insurer already has an asset, such as an efficient administration platform that they can offer and tailor it for a fee or lease it out for a fee. And then there are value-added services such as digital services out of these data insights that these interests hold. So, there are a lot of prizes behind this.
And to come to the application of it to commercial insurance, I would like to hand it over to Sarah.
SARAH STEIGER: Thank you very much, Silke. So, what does this all mean for commercial insurance? We clustered some of our research into four topics and we can go through them and elaborate on what they are stating. So, the commercial insurance market, especially when we talk about the small commercial insurance market, is a rather fragmented market. We have a lot of large players who take up a large part of the industry, so around 30-40%, and this gives them the advantage to change the market scenario as they can scale and apply technology to these changes. However, the established commercial insurance companies often struggle with legacy systems and services and this is challenging for them to make the transition to the new.
We have on the other side agile new entrants that can enter the market and they can bring customized and more simple products to the market, and this allows them to make an impact with more customer-focused services and easy to access digital direct distribution strategies. We have further identified, when talking about distribution, that a large part of the expenses the insurance market embraces are related to broker fees. So, the broker charges account for 26% of the fees paid by insurance companies. This is for the US P&C market. Brokers, in turn, generate leads for insurance companies and secure an in-flow of business and renewal season, as well. So, it is believed that for the small business market, the agents and brokers will continue to play a crucial role in this setting. The question, however, is how the role of a broker will evolve with all these new agile new entrants coming into the market, who are creating these direct digital touch points with the customers.
When we talk about products, and as we have already mentioned earlier, our research shows that new risks are evolving, and new exposures are emerging. We see that the market is witnessing an increased demand for products that address specific exposures, which as Silke mentioned before, we think about the cyber exposures, exposures related to the gig-economy, and pandemic exposures, et cetera, just to name a few. So, here is where the question is posed, how can you react as a commercial insurance carrier to these exposures to price them adequately and cover the risk that is emerging with these new exposures?
And last but not least, I’d like to talk about underwriting briefly. Underwriting lies at the core of the insurance value chain. With the continuously increased availability of quality data in the commercial sector, the analytical ability of the automated underwriting process is also increased. So, we can increase straight through processes. And with this increased automation, underwriters are freed from the repetitive tasks they might have had to continue before. A lot of companies still work with Excel datasheets, as we saw earlier, these come from the early 90s, and now the underwriters can focus on value-adding tasks in their daily jobs. And so, the digitalization of these underwriting processes is also a key step towards augmenting of the direct distribution capabilities of commercial insurance carriers.
So, what we’ve put up here is the insurance value chain and we have selected a few, to show you a wide array of different selected use cases that we have identified along the value chain. This is just to give you a small flavor of what is possible. Within this, we have highlighted two, within Marketing, Sales, and Distribution. The first one is the lead generation and prioritization for agents and underwriters. The benefits of this use case obviously include increased efficiency of new business and lead generation for the commercial insurance sector. The second use case we would like to highlight is Next Best Action, which we’ve already briefly touched upon before. Here the actionable recommendations are coupled with a 360-view of your customers via a client cockpit, and this leads to increased service quality. These are just a few of many use cases which can be applied to the insurance industry. And with this, I would like to hand it over to Miguel who will introduce the AI-driven client cockpit. Thank you.
MIGUEL RODRIGUEZ: Thank you very much, Sarah. So, all the use cases that we see here are use cases that are out there in the market, we will now focus on just two of them. All the others are applications, cases that can be covered through our platform and through our apps. So, I’m happy to discuss those as well in more detail separately.
So, if we look at the first part, we have here Squirro’s Client Cockpit. What do we mean when we talk about the Client Cockpit? The Client Cockpit with a 360-degree view is not a new topic, it has been on the table for quite some time. But in most of the cases, the Client Cockpit was always around gathering data together from internal structured sources. Less around unstructured data sources, from internal systems and internal data silos, also from external data providers, opinion data providers. So, the first step of what we define as the Client Cockpit is to get data from multiple internal and external data sources, and not only taking into consideration the structured piece, but also everything that is text and unstructured and siloed, and hard to analyze. This is then the data that we need to comingle, the structured and unstructured data, for a holistic 360-degrees.
With the second step, just bringing data together is not enough, because just more data, not in your context, is not interesting. That’s why we integrate the Client Cockpit within, as an example, a CRM system like we see here. It could also be other applications, to make the insights available to people where they work in a seamless way, as we had discussed earlier in the webinar. To make this happen and make it more tangible, we need to filter out the irrelevant information. So, we do this by applying Squirro Score Capabilities around unstructured data analytics and we also detect catalysts within this data that can mean a direct call to action for the relationship manager. So, you would understand if a certain event has happened, see it is directly in context of the account, and will be ready to make the call directly without having to go for the information. And there is a part of that, as well, if you look at Next Best Action recommendation, in what will be the right product to offer to a prospect. As you can see here, and this is just a screenshot, everything is put in context of the client, everything is made available right people work. Everything is filtered and put in context, analyzed, and the insights are presented in easy to use and easy to understand dashboards so that data analytics, AI, and machine learning doesn’t remain a topic that is only available to people that have advanced skills and can work with the solutions. But, it’s something that you can make available for everybody within the company. This example is now more for the front of the company, the same applies also for the underwriting process, for customer service, for claims, et cetera.
If you look at it more on the deal origination topic, I think that is a very, very important one. We heard before that a lot of costs in the commercial insurance base goes to brokers because they bring business and insurance providers must pay for it. And here we think that there is great potential for commercial insurances to generate additional revenue. Because a lot of information is out there, it can be within annual reports, it can be within earnings call transcripts, it can even be within call notes, emails, document, et cetera. It’s not used today, because there is so much information and usually the underwriters and the relationship managers are very busy with day to day work, that it would require too much time to go through all this information, find the insights in there, the really valuable nuggets, to take an assessment and understand when they could call someone and when not to.
So, first, the important part is when everything is gathered together. As I explained before, when you apply data analytics to it, you use this to provide a deeper understanding of the client and industry, get more insights, and have real-time identification of what is going on in the market. Then we have our pre-trained AI models that are able to detect any unstructured data, such as earning calls transcripts, which are usually very long, or other events that are very relevant, for a relationship manager. This comes out of the box, pre-trained, ready to be applied. And our application is capable of reading line by line through the earning calls transcripts, for example, identify the catalysts, highlight the sentence that points out, as an example, that the company plans to invest into machinery and technologies, and also the amount that they plan to invest. And then we can provide the direct recommendation within the CRM system, as we see here with Optotune, to the relationship manager, that his company is planning to do this investment. So, if they plan to invest in machinery, as we see here, obviously they will need to have insurance for it. It’s a great opportunity for the relationship manager to take the phone, give a call to representatives at Optotune, and have a direct point of conversation saying, “we read in your annual report that your CEO states that you plan to invest in machinery, have you already thought about how you want to ensure those machines?”
So, you have a direct lead and you can follow up without having to go through all these different data sources and data silos and spend a lot of time there. And as we know from the many implementations that we did and customer feedback that we received, in 80% of the cases the one to call first is the one that makes the deal. So, time is very important, and information advantage, obviously, as well. So, we see here a lot of benefits for the users of this commercial insurance application that combines a lead generation element, the next best action element, together with the client cockpit. The benefits are really increased efficiency, helping the business, and lead generation. We have a more proactive approach to new opportunities with the analysis of not only the structured data but also the unstructured data. We have deeper insights and better customer understanding. We can improve customer service and increase the client’s satisfaction.
And with this, I will end the webinar presentation and we will now go into Q&A that will be moderated by Eric.
ERIC ROBUSTELLI: Great, thanks Miguel for that. Thank you, Silke and Sarah, as well. That was some really insightful and thought-provoking conversation we had there. So, we do have a few questions from the audience. The first one goes to you, Miguel. “Can you please give some concrete examples of the value-added by data insights?
MIGUEL RODRIGUEZ: I think we covered, quite a bit already, the advantages that we can provide. So, on one side, using data insights in a B2B situation, it provides a lot more information to have a better, more informed conversation with your client. You’re always up to date on what is happening. It gives you an opportunity to optimize meeting preparation times, as everything is available in real-time and in your context and context of the client. You can start the research right before the client meeting. And then obviously if you consider also the lead recommendations, the data insight provides a lot of new leads and new opportunities to engage with customers that were not there before, as we transformed from maybe a bit reactive approach, waiting for a broker, to a practical approach that is led by data insights.
ERIC ROBUSTELLI: Great, thank you. Silke, this next one is for you. “What, in your opinion, are the main disruptors to the insurance value chain at the moment?”
SILKE GENUIT: Yes, we’ve mentioned before, there is a lot of new digital players out there, new insurtech companies, also the GAFAs, even companies like Ikea, or Amazon, especially, who have lots of customers, lots of clients, and access to lots of data, are maybe not yet in the insurance business but maybe are thinking about moving into it, to leverage the data and to leverage the insights they have on these customers, to offer exactly these personalized insurance covers, essentially covering the right channels, at the right time, at the right place to their clients. So, there is a lot, I think, out there and it will be interesting times.
ERIC ROBUSTELLI: Great, thanks. Miguel, another one for you. “Which use case represents the low hanging fruits for commercial insurers?”
MIGUEL RODRIGUEZ: That’s a very interesting question. It really depends on what’s the main pain point the insurance has at the moment. We have tackled topics that are simple, without underestimating it, like searches, combining the data that they have internally to provide them faster access to data that usually is siloed and hard to access. We obviously worked on client cockpit use, I think that’s also something that can be done quite quickly, and specifically the comingling of the external data with internal data. And having pre-trained models like the ones that we have, that detect those catalysts, those events, automatically, is something that can be deployed quite quickly and provides direct benefits to the insurance company from day one.
ERIC ROBUSTELLI: Thanks, Miguel. So, another question for both of you, Silke and Miguel. “How do we prepare the workforce to embrace and not resist AI?”
MIGUEL RODRIGUEZ: I think that a very important piece is the approach that we choose. A lot of companies out there are also talking just about AI and robotic process automation and a lot of topics. We talk more about Augmented Intelligence, and that implies directly that we augment the humans, we help the human in day-to-day work. That’s why our applications, our hope is to provide information in context of a system where people are actually working. So it’s not a replacement solution that we have here, but it’s an enhancement solution that can work with other tools that people use in day-to-day work. Making AI understandable and seamless, the same way as we saw it in the retail space, I think is the right approach to make it also usable for people and make them embrace it.
ERIC ROBUSTELLI: Silke, any thoughts?
SILKE GENUIT: Yeah, I was just going to add, I absolutely agree and we think it’s important to support, as well, the up-skilling of the employees, to have the right training, have the right understanding of where the AI-related technologies help them, support them, and is important their work and output and help them in any way. So, I think that up-skilling is an important factor. Also, for the companies who look into using Augmented Intelligence, it’s very important to first know what their actual strategic goals are, where do they want to go, where do they want to be, how do they want to get there, and look at it from that perspective, not just looking at single or simple use cases, but overall think about what are your strategic priorities, how can technology help you reach these, how does this all come together, and where can you actually best leverage the technologies that are out there.
ERIC ROBUSTELLI: Great, thank you for that insight. So, we’re coming to the end of the webinar. We did receive some other questions from the audience, so with that said, we are going to be following up via email for the rest of the questions. We will also be making the recording of this webinar available for everybody who joined and everybody who registered, as well. And with that, thank you all very much for attending Squirro and Accenture’s webinar here today. And have a great rest of the day.
Thursday, May 23, 2019
10:00 am EDT
03:00 pm GMT
04:00 pm CET