Page 1


Reinventing the wheel is expensive and risky

This white paper considers a dilemma which many businesses have had to contend with, namely, the question whether it is better to develop the technology you require for your business operations yourself or to obtain it elsewhere. This is a question facing all businesses and the answer is likely to differ from one industry to another. In this white paper we focus on automation and digitalization within the insurance sector. Within this sector we consider the supply of analytical solutions for automated risk management and fraud detection.


IS BUYING BETTER THAN BUILDING? Let us start with the buzz words. Many companies are considering the use of AI, machine learning, data mining and applied analytics to improve the customer journey and experience, two elements that are more important now than ever before. How can these technologies be employed in the current system topology? And where do we find the expertise and experience to deal with them? Let us begin with a conclusion. For most insurance companies, the optimal solution lies in the utilization of specialist suppliers. In this paper, we explain why this is the case. The aim is to provide the reader with questions which they might ask when such matters are considered within their own company, allowing them to draw the most appropriate conclusion. For this purpose, we provide a checklist which the reader can use to make a comparison between Do It Yourself (DIY) or buying a solution for fraud detection and risk assessment.

WHAT IS THE RATIONALE? There may be various reasons to automate your current processes. In some cases, it is necessary to switch to a modern core system. In other cases, all that is required is to upgrade or modernize core system components. The aim is to ensure that the organization’s immediate channel can develop within a rapidly changing market, or to enable its portfolio to grow sustainably without taking unnecessary risks. It is often those who are combating fraud who indicate that an automated fraud detection system could significantly boost investigative outcomes. Underwriters may also want to increase their pace and become more customer friendly by making prudent decisions at lightning speed based on real-time data analysis, providing their customers with the quickest and safest service possible.

BUILD OR BUY? THAT’S THE QUESTION. Whatever the reason, when there is a need for modernization, the question quickly arises whether a company is going to develop this themselves (or arrange for it to be done) or procure it elsewhere. In the past, systems were often developed in-house. The IT department was assigned the responsibility to investigate what was required, create a design, and develop a resource plan, determining which staff skills they could provide in-house and which skill sets needed to be filled by external staff. The IT department was then also assigned responsibility for the maintenance of the new systems. This build process made sense, due to a lack of specialist systems . The mindset was that system requirements were best understood by the people working within the organization. Perhaps there was a fear that market or business secrets would not be secure.


INSURTECH SPECIALIST The situation is different now. Highly specialized technology providers with in-depth knowledge of the insurance market have developed systems which address specific needs, such as systems geared towards risk management and fraud detection. It is safe to say that due to a technology provider’s expertise and knowledge, it is virtually impossible for a non-specialist developer to achieve the same quality and innovation. The specialist focuses on offering the best, most effective automated risk management and fraud detection system for today and the future. They do this by constantly remaining in touch with their customers and industry analysts. They monitor customer requirements and work with the analysts on the latest developments in technology and data collection. They continuously add new features/functions based on the feedback loop. An insurance company’s core business lies in offering optimum financial protection and not in the development and continuous upgrading of automated solutions. This is what an insurtech specialist does.

WILL WE STILL BE ABLE TO DISTINGUISH OURSELVES? Insurers wish to distinguish themselves. This may be done in various ways, through their image, customer experience, or a combination of both. Sometimes insurers fear that if systems are provided by external technology provider specialists, they will no longer be able to distinguish themselves from competitors. This is sometimes a reason for companies to take matters into their own hands, based on their own corporate strategy and “risk appetite”. Data analysis systems are flexible and may be fully configured in accordance with a client’s wishes. For instance, the sources that are to be used may be expanded or limited, and the outcomes produced by the algorithms may be adapted to such an extent that they are aligned with a company’s corporate strategy. In this case, the system is not an IT solution which is managed on the basis of IT expertise. It is a tool that is directly related to corporate policy and which may be adapted immediately to accommodate any modification, without the need for a single IT assistant to be involved within the company.


WILL IT FIT IN WITH OUR EXISTING SYSTEM? Automation and digitalization used to be processes of incorporating individual new features into a pre-existing – usually on-site – system. This process produced large, often overmanned IT departments that were required to keep everything running. Every new addition demanded a considerable investment of both time and money. This situation is changing rapidly. With the adoption of cloud, no matter how complex or dated an existing system is, a risk management and fraud detection system can be integrated into a company’s core system without any far-reaching modifications. While this operation is easier to execute in a secure cloud environment, it can also be integrated with on-premise systems.

WILL EXISTING SOLUTIONS BE ABANDONED? Some insurers developed their own risk management or fraud detection system in the past. Theses systems already have connections with data sources, such as the Chamber of Commerce, the Personal Records Database, the motor vehicle registry, sanction lists and sources within the company, such as its claims administration office and the registers of the special investigations unit. Frequently, specific business rules are already applied in the acceptance processes and for the purposes of claims handling. Those solutions will not be lost if a new, automated data analysis system is introduced. In most cases, they can be incorporated into the new system.

Case study Insurer X detected 30 suspicious claims on average per month using its in-house developed data analysis system. Every year this reduced the # of wrongly processed claims, saving the company a significant amount money. After an insurtech specialist conducted a study of this area and prepared a business case, it became clear that further improvement was possible. An automated fraud detection system was implemented. The number of suspicious claims that were immediately detected rose considerably and the number of “false positives” declined. As a result, the initial investment in the system was recouped in its entirety within three months.

DO I HAVE THE EXPERTISE IN-HOUSE? Not only does developing a system take a great deal of time, it also demands specific expertise, for example, when it comes to data analysis and algorithms, as well as privacy and compliance. Some large companies may have such expertise in-house but most will need to hire it. This is not only expensive but also risky, because any expertise that is hired temporarily will probably no longer be available at a later stage. To an insurtech specialist active in the field, this is their day-to-day job. These specialists always possess the latest knowledge and will ensure that the technical aspects of systems are current and in compliance.


WOULD YOU LIKE TO KNOW IN ADVANCE WHEN YOUR SYSTEM WILL BE IN WORKING CONDITION? Of course it is possible to build your own data analysis system. You may succeed in doing so, if you manage to find the appropriate specialists, and provide that the most appropriate design from the very beginning and you are willing to invest a great deal of time and money. In addition, you need to be prepared to commit to ongoing maintenance and upgrades well into the future. Usually, this is not the reality. A large amount of time and money is required for the design and build, often far exceeding what was originally budgeted for this type of project. The reason for this is that you are essentially reinventing the wheel and will encounter a range of obstacles. This starts with internal capacity. IT departments are often busy (or too busy) with maintenance and projects. A data analysis project gets added to the bottom of the list. In addition, the requisite expertise is often not available in sufficient amounts. The situation is entirely different in the case of an insurtech specialist. You are immediately at the top of their list for implementation, and they have the expertise. The system has already been developed, is being maintained with the latest insights from all implementations, and can be fully operational within a relatively short period of time - always much sooner than if you were to build it yourself.

WOULD YOU LIKE TO ALWAYS HAVE THE BEST SYSTEM FOR THE BEST PRICE? It goes without saying and we have mentioned it before: the specialist always leads the way. Developments in the field of automated data analysis happen so quickly that it is impossible for a non-specialist to keep up. You can fall behind in the blink of an eye and for most insurance companies it is impossible to constantly implement innovations in their systems. They simply do not have the capacity and expertise for this, not to mention the budget. By their nature, a specialist always wants to move ahead and is also capable of doing so. A specialist operates in this market day in and day out, maintains contact with a number of clients, monitors all technological developments and constantly introduces innovations. Their clients can benefit from all of this immediately, without needing to make any additional effort themselves or to incur unforeseen expenses. Any company building their own system would therefore probably fall behind very quickly without being able to catch up and will also lag behind in terms of profit. Companies that purchase their IT solutions always have access to the best systems, without uncertainty about the costs involved. The costs are clearly stated, which means that there are no unpleasant surprises and companies can maximize their profit.


IS BUYING MORE AFFORDABLE THAN BUILDING? While there are exceptions, nine out of ten times the answer to this question is an unequivocal “Yes”. When building a risk management and fraud detection system, you usually work from the bottom up. The process of designing, developing, testing and adjusting takes a great deal of time, in some cases years. This is because the project needs to be incorporated into the IT department’s already busy schedule. Postponements and delays are constantly in the offing. The situation is different when you buy a system. In this case, you are dealing with proven technology, which is usually up and running within a few months. It yields results from the outset and you always know exactly what costs are involved.

SAAS OR ON-PREMISIS? If you opt for buying, you then have the choice between a software as a service (SaaS) and an on-premises solution. Many choose SaaS. In the case of SaaS, you do not buy any IT but rather a specialist function, which runs entirely outside of your company and is constantly maintained and expanded by specialists. Every new feature of automated risk management or fraud detection is implemented and made available promptly, and maintenance occurs throughout the year. The system always complies with the relevant compliance and security requirements. If you opt for an on-premises solution – where a technology provider builds it within their client’s system – an equally good outcome can be achieved. However, the implementation period is usually longer and more complicated. In addition, maintenance depends on one’s own IT organization and every improvement must always be installed separately. The costs involved are usually higher and less predictable.

WOULD YOU LIKE TO KNOW WHETHER YOU CAN ACHIEVE A PROFIT ON COMPUTERISED DATA ANALYSIS? The implementation of an automated fraud detection solution is a major decision you shouldn’t take lightly. You would like to know beforehand how much it will cost and what it will yield. You would also like to know what efforts you will need to make to implement it and how much time that will take. In order to provide this information, the technology provider often creates a business case together with their client. Based on an analysis of their client’s requirements, their corporate strategy, technological environment and all available data analysis systems, a description is produced for the impact a new system is expected to have, its anticipated return and the break-even period.

COMPARE FOR YOURSELF If you still have any doubts as to whether you should buy a system or build, we recommend that you use the list of decisive features below. This list will help you in the decision-making process.


DIY Price of licence Price of maintenance SaaS On-site Claims fraud analysis Underwriting risk assessment SIU investigation support Compliance coverage UBO detection Personal lines coverage Commercial lines coverage Case management Internal insurer data Proprietary data Third-party data Fraud pattern recognition Voice analysis Social network analysis Predictive analytics Time analysis Run-time module scheduling Reusable variables and models Ability to compare scenarios Anomaly detection Text mining Image screening ASP Ability to design dashboards Ability to share fraud cases Ability to prioritise cases Ability to improve the system by running models Audit trails provided Time needed before going live PEP lists included Subject matter experts Full service delivery Self-service delivery Mixed model delivery

Vendor 1

Vendor 2

Vendor 3


CONCLUSION Remember that building an automated fraud detection and risk management solution is like building a house; it’s an investment that needs the right mix of expertise, specialized tools and skill sets. The good news is that there are automated fraud detection solutions to help you make a more thorough and worthwhile Buy vs. Build decision.

Would you like to know more about FRISS? Orteliuslaan 15, 3528 BA Utrecht, The Netherlands

(+31) 30 767 0352 marketing@friss.com


Profile for Maarten Uri

Whitepaper: Fraud detection: The build versus buy debate  

Whitepaper geschreven door Maarten Uri in opdracht van FRISS Fraud, Risk & Compliance www.uricom.nl

Whitepaper: Fraud detection: The build versus buy debate  

Whitepaper geschreven door Maarten Uri in opdracht van FRISS Fraud, Risk & Compliance www.uricom.nl