2018 was marked by numbers of severe and ubiquitous frauds and abuses. It has already been a while since fraudsters are no longer hooded teenagers in their mother's basement. It is a professionalised activity with a lot of resources.
We have spent the year reaching out to new industries and showing the benefits of AI-driven fraud prevention solutions: from regional digital lending enterprises to the space industry without forgetting airlines, online travel agencies, and luxury retail companies. With the new year quickly approaching, it is time to reflect on all the work we have accomplished and what still remains to be done.
Unlocking a new market: Digital Lending
Fast-growing industries tend to attract a lot of fraudsters. With the ever-growing automation of digital services, the financial services industry was one of the most victimised ones. It is not surprising as financial services typically deal with payments that reduce the cost of money laundering for fraudsters. Especially, if compared to the purchase of luxury goods or plane tickets.
We got contacted by these businesses at the beginning of the year and have taken it upon ourselves to understand the intricacies of the business and build personalized solutions for this particular vertical. We have had some big successes:
It remains the key capability Nethone has built over the last 4 years. Digital lending companies, no matter how big faces this issue and we have been able to stop thousands of fraudulent loan attempts for our clients in Poland, Spain, the UK, and Brazil.
Cost reduction models
In this vertical, risk assessment is important. This is why all of the digital lending companies spend a small fortune on credit scoring agencies which provide them invaluable information about the capacity of repayment of their potential customers. We have been using all the power of Nethone to create models kicking out users with very low chances of getting past the credit scoring stage before these companies ask for financial data. It has helped our customers to lower their variable operating costs by 20% while keeping their overall conversion rates at the same level.
Credit scoring for emerging markets
There is a huge interest in digital lending for countries with little access to credit. Typically, these countries have large populations of people with no bank account. It makes measuring their risk of default very difficult since no data is available about them. Nethone has been developing technology gathering over 5000 data points about each user solely through their current session (with no external data sources used). We have been working together with our clients to help them expand to these emerging markets. At the same time we were keeping default risks in check thanks to these data points.
We would like to take this opportunity to thank both our clients and The Polish Association of Lending Institutions. Without them, we would not have been able to create all of this value. We are looking forward to bringing more innovative solutions in 2019.
Great leaps in the travel space
Nethone payment fraud prevention solution, Nethone Guard, was born in the travel sector in Latin America 3 years ago. We know this space well, yet, this year we have noticed that we still have a lot to learn and have been lucky to be working with the right partners to continue improving.
Through the expansion of our current partners and the acquisition of new clients, we have been exposed to 5 times more traffic over the last 12 months. Not only have we seen more traffic but we it was also coming from more geographies. It has been essential to the continued development of Nethone Guard. Below some of our biggest steps forward:
Fraud attack prevention
No matter how well a system is set up when a new type of fraud comes along there is nothing that a system can do. Or can it? We see everything in the context of data: if our machine learning engines have not seen a type of fraud before they could not have learnt to prevent it. However, we noticed that attacks always come in a concentrated form. This is why velocity rules exist - we decided to create a "fuzzy" velocity rule which checks all 5000 data points we gather for "density" in the traffic. If too many similar transactions came into the system, we would warn fraud prevention teams about them or even better: block them. Promise, no more bragging! We are just really proud of this new ML engine.
Foundation models for new geographies
We are seeing more and more traffic from new sources. It has helped us to develop "foundation models" which are simply models applicable to new client's data for new geographies. It has allowed Nethone to forego the need for historical data for new clients and allowed us to give high-quality recommendations from day one to our new clients in chosen geographies.
Ready, steady, go! E-sports and ATO
Doping in e-sports damages the reputation of the players and deters investors who have an appetite for a piece of this profitable pie. In January, Nethone launched two extensive R&D projects subsidized by Poland’s National Centre for Research and Development (NCRD): Nethone ATO and Elympics. The total value of the two projects is $5.98 million (including $4.45 million NCRD subsidy). We have taken an integrated approach: looking at each player's behaviour, skills and history to connect with rivals with comparable parameters and automatically detect the symptoms of e-doping.
The aim is for game developers or studios to implement the tool in their products in order to both increase mobile e-sport security, but also significantly reduce the cost of game development — critical for this industry growing at a very quick pace.
Speaking of growth, account takeover has become a thorn in the flesh of quite a few companies. Successful ATO prevention is more than just two-factor security, fraud monitoring systems, or daily transaction limits. ATO done properly should be frictionless and allow marketing managers to continue improving conversion rates without having to think of security implications. Nethone ATO is the combination of real-time protection with undisrupted UX. Launched in January, the project is an autonomous AI-based cybersecurity solution aiming at quick, user-friendly protection and insights into account takeover attacks of financial institutions.
Some of the biggest data breaches and hacks in history have happened within the last 12 months. Cybersecurity should be at the top of every corporation's mind. Nethone ATO capabilities allow us to go through every one of your users and assess their hardware, software and behavioural profiles for risk. Using Machine Learning, we are able to take into account over 5 000 raw data points before making a decision about a given user.
Looking back at 2018 we see the growth, accomplishments and many lessons learnt. From doubling the team of Nethonians to the business development achievements - it was yet another fulfilling year. We’re looking forward to more challenges in 2019: this year shows we are more than ready for them.