Before talking about technology in the insurance market, it is important to note that one of the consequences of the pandemic was the recategorization of priorities: Before the pandemic, we saw carriers implementing innovative technology to fight fraud, because fraudsters did not use regular channels, but rather technology to commit fraud, and so it called for a technology-based response.
Then the pandemic took place. In addition to the shifts in technology, in each country or region, many carriers have undergone a reprioritization process, and this often has a negative impact. Technological innovation has either stopped or shifted in the short term. And why is that happening? One reason, for example, is that we were not ready to work remotely. We need to figure out how to be always close to our customer. And, obviously, this customer has also been affected by the pandemic, so they themselves are starting to use new ways to commit fraud due to a shortage of income or job loss. In fact, there is a combination of factors: on the one hand the carriers and their shift of priorities and, on the other hand, the fact that we were not ready.
Fraudsters also use technology
During the pandemic there was no physical contact, and many claims were submitted, or potential customers sent requests through a computer, and the carrier had no way of knowing if they were talking to the person they claimed to be. So, there was no way of knowing if the information they were providing was accurate. Many fraudsters use this to their advantage and, obviously, both policyholders and intermediaries were faced with an almost exponential increase of this very frequent broker’s scenario: “how can I make my carrier believe in X when the reality is actually Z?” So, fraud committed either by end users or internal users (internal fraud) alike is what many carriers have been recently experiencing.
Which is the technology that is contributing the most to fight fraud?
In general, it is best to consider integrating all technologies to get the most of it, but let me explain what I mean. When a carrier just starts, they need a lot of data to run predictive analysis and to drive artificial intelligence, and they cannot start from just a trend, derived from an incident, to develop a model because there is a lot of data missing. So, carriers face the challenge of integrating different technologies. Take the case of telematics, knowing the user’s behavior, whether it is through geolocalization, or what their behavior is related to speed, and integrating that data to my analysis can tell me a lot more about my customer than just funding out they are doing something wrong. To us detecting fraud already sounds like something negative and what we want is to focus on honest customers, to provide them with better services, to make them move forward. That is the goal of the carrier rather than tracking down bad customers. The idea is integrating various technologies to the carrier’s processes. To many carriers it is much more valuable to know if their policyholders have had some legal issues rather than what was posted on Facebook or other social networks. Some may say, I need to know how many fines my policyholder has and that is why we have a data provider. Others will focus on where the customers are geolocated, where they live, and within that process each carrier will have to assess, together with the broker, what impacts the most at underwriting.
Image scanning
What ML (Machine Learning) does is to ”learn from situations”: from pre-existing damages, from vehicle manufacturing, year, make, all those details are stored, and the system learns more and more. So, when a potential customer arrives or a claim is submitted, we use this technology to get a result which is then processed by FRISS predictive analysis and this output is added to the existing investigation. This technology exists and is being debugged. We have a customer in Germany who uses it at underwriting and not so much with claims.
How does this technology help assess risk at underwriting?
There are two cases: it helps at the time of quoting and when issuing a policy. First (at quoting) we have basic, general data. The carrier does not know if I am in the United States, Argentina or Singapore. It only needs basic data because I am just adding items to my shopping cart. I am not going to use a lot of technology or external data at the time. This bears a cost to the carrier, and I don’t know whether the person will actually buy the insurance. So, at the time of quoting, carriers make few checks to issue a quote. There are standard prices at the marketplace, but in the end, I want to get a new customer, be it through a digital platform or a broker.
Then, when issuing a policy, every item must be checked, which is why technology plays an essential role. When the carrier issues a policy to Mr. John Doe, that person and the object to be insured must comply with company standards: liabilities, job position, income, etc. That is, if the income can cover for the policy, if they have historical records inside or outside the company, and if what they tell us about the object is true, the item which now belongs to Mr. John Doe, but was previously owned by others. All the above is connected to technology: checking the data and getting information in real time. And that is what we do at FRISS. We provide that information. We then connect the object in question with the relevant person and see whether it makes sense to run predictive analysis to say, in real time: “Hey, we accept or reject the policy” and we customize the quote. So, technology is important at underwriting. It is hard to say this but the world of underwriting and that of claims are often quite different. We need to be clear about this. We need to be able to check claims to decide whether I can approve a policy. Technology helps us have a healthy portfolio
Occupational hazards — Technology that helps detect fraud
There are two sides. There is an a posteriori situation when someone files an occupational hazard case, and such person reports , for example, that they broke a leg, but on Facebook we see that they are doing this or that thing instead. Such an investigation is costly because there are man-hours devoted to trace the case and debunk it. And there is also an a priori situation where the carrier cannot do much at underwriting and they need to use technology to drive in much more data. Especially in the United States, we have seen they know well in advance these standards we were talking about: physical and mental work, and the type of known risks to make an assessment. Making measurements with everything they know so far to create a model. In any case, people who are determined to commit fraud end up doing just that. We have heard many times: “I was going to my work and suddenly I got into an accident.” Another interesting thing that goes beyond technology are the conditional clauses people must comply with. We cover for the policies only under these conditional clauses. But this is often not clear to the user. They think: “I am insured, and my policy covers for all risks.” And they often get maimed, or they intentionally get hurt, and it is necessary for the carrier to consider such situations in order to prevent the risk. This is a bit more difficult because we need more data in this case. So, the carrier needs to collect that information which they do not have, and they need to create a scenario. In that sense we can say: “Hey! There are no medical records, but you can create a model, we can put together something with the information you have,” and gradually data builds up.
New customer SURA URUGUAY SERVICIOS
We have a corporate-level partnership with SURA Uruguay, part of the SURA group. We first noticed some issues: a low combined ratio and some processes still manually run. The truth is that this is part of a larger program that involves innovation, not only for SURA Uruguay, but at a regional level, a very comprehensive program that goes beyond fraud prevention, and thrives on the technology that really works for them. At SURA URUGUAY SERVICIOS we currently implemented two solutions: One for claims and another for Investigations. The latter is a large differential FRISS has compared to other suppliers: It is a completely separate and independent module, and it can be fully integrated to all other FRISS solutions. It helps carriers that do not have any other FRISS solution in place. This means they have the entire workflow of an investigation; you don’t have to digitalize anything because the information comes from Claims. If you want to manually add a claim, you may enter types of amounts, specific to that carrier, connect to a fraud pool if there is one, give access to external investigators, manage an investigation, upload audios and videos. We also have image screening, that is, we use the photographs sent to us to check if it has been downloaded from a server, etc. A large number of features all integrated into one module which makes it much simpler and faster to fight fraud.