56.4 F
New York

Volodymyr Todurov on aI-driven anti-fraud solutions for PIN-UP GLOBAL

The iGaming market is on the rise, and such rapid growth occurred due to implementation of the latest technology. Businesses can offer customers top-notch products and services, but this is possible only if they use the latest technology and approaches, from High Load and Big Data to High Scalability. PIN-UP.TECH specializes in taking risk management to new heights and creating top software solutions that may change the industry in the best sense possible.

The chief analytics officer, Volodymyr Todurov, commented on how the team evaluates the role of innovations, ML, and AI, and shared his approaches, so read on for more insights.

The basics: Key information on PIN-UP

PIN-UP.TECH is part of the PIN-UP GLOBAL holding. It consists of multiple IT companies and creates unique solutions for partners to achieve new heights in the iGaming market. Now the team comprises of over 500 multinational experts and specialists focused entirely on the stability of the platforms, CRM, anti-fraud solutions, affiliate systems, and more. Volodymyr Todurov is the Chief Analytics Officer at PIN-UP.TECH, and he mentioned many interesting aspects of risk management in his recent interview – https://sigma.world/news/volodymyr-todurov-about-ai-anti-fraud-usage-within-pin-up/.

Volodymyr Todurov on antifraud tools based on ML & AI

Some of the most important parameters to consider are the history of bets, the size of bets, and the frequency of bets. AI models can use them all for in-depth analysis and focus most on the criteria with higher predicting power to make more accurate forecasts.

Todurov says that their team concentrated mostly on the extreme gradient boosting model that is proven to be very effective for structure data. This type of system allows the creation of an entire set of models that are constantly improving one another and increasing the overall accuracy of forecasts. You basically share a lot of information — the betting history with it, and it yields a probability of 0 to 1. This rate indicates how likely a particular user is to engage in fraudulent activity.

Volodymyr Todurov on analytics based on aI: Key benefits of advanced technology

Volodymyr Todurov says, that ML and AI models are just irreplaceable when it comes to analysis of large amounts of information. They are better at generalizing information and spotting even non-obvious behavioral patters. Manual transaction analysis or linear rules, Todurov says, will never give you the same or even similar results.

Still, though technology matters, we can’t say it will replace human intelligence. The team of PIN-UP.TECH uses AI and ML as additional tools for optimization. In the end, these are the models that help reduce errors and spend less time on decision-making. Moreover, it has an interesting “side effect”. Specialists who use artificial intelligence can broaden their own horizons and perspectives by taking a new look at the problem, feature, criterion, etc. AI can already complete part of the tasks fully independently.

All in all, you can do more without hiring more people and remove a more routine component focusing more on other important processes that haven’t been addressed previously.

Volodymyr todurov on anti-fraud platforms’ future

Volodymyr and the team clearly plan to use machine learning and artificial technologies further. In particular, PIN-UP.TECH is going to create a decision-making model using deep neural networks and generative AI. That would allow us to make decisions faster, Volodymyr says, and in the end, we’ll get way more accurate predictions about fraudulent activities.

The best thing about this scheme is that the model will be evolving. It’s created to learn and adapt, analyze more parameters of customers’ history and identify separate customer segments, including the fraudulent segment.

Businesses, in turn, can do retrospective checks. That’s how the teams of operators can ensure that the client is labeled as “reliable” or a “potential fraudster” correctly. Managers will feel less pressure because they won’t have to worry that they’ve missed a suspicious account. AI is here for manual tasks. Needless to say, that greatly improves the effectiveness.

Volodymyr Todurov on risk management & anti-fraud advancements

As noted by Volodymyr Todurov, the PIN-UP team has its own vision of how the holding can optimize work and cope with potential difficulties faster and more easily. The main focus is automatization and a data-driven decision-making approach. The team analyzes its own workflow, aiming to spot things that can and should be optimized.

For example, Todurov notes that they used to export a few CSV-tables to see whether there are any behavioral changes on the customer’s accounts, focusing mostly on the history of transactions and bets made.

In most cases, such manual analysis took 10-15 minutes. That’s why we decided to create a default real-time dashboard with all the charts and tables.

Here’s another example. Of course, we monitor financial transactions — withdrawal confirmation is an essential part of the work. The above-mentioned DDDM approach is used to analyze the predictive power of specific evidence and then manually confirm the withdrawal.

We identified certain parameters based on previous fraud cases to spot repetitive scenarios, Todurov says. However, the business was growing, and more users from different locations were joining the community, so the scheme wasn’t working as well anymore. That’s why we decided to adopt the rule, and that allowed us to boost our overall productivity by 15%. So, the team’s goal is to constantly monitor the situation and analyze all the old and new factors to optimize and automate all the processes.

DDDM Model integration: Speed, challenges, and potential benefits

As noted by Volodymyr Todurov, the iGaming antifraud DDDM approach requires a certain mindset. Even senior managers with extensive experience should refuse all decisions based on gut feelings, expectations, hopes, and “expertise.” Todurov says that he realizes such an approach may still be used in exceptional cases when there is no data or when the problem isn’t really worth spending too much time on. But when you decide to change something in your product, you have to ask specialists for specific figures.

Specialists should also try to get rid of cognitive distortions, in particular, the survivorship bias. Decisions should be made only based on end-to-end analysis. Todurov says that they had a case when a manager came with an analysis of fraudulent behavior, bonus abuse, to be exact. The analysis was carried out within a 4-month frame and has shown that 83% of clients had the same patterns: made a first deposit of $13+, used a pre-paid card, and had an IP mismatch due to proxies or VPN. However, though the portrait is quite detailed, the calculations showed that if we apply those parameters to new customers, the total efficiency will be just 5%. It would affect larger groups of customers with similar choices yet no fraudulent behavior.

If you focus on the DDDM approach, you will definitely get some great results from the very first day. You’ll reduce errors and boost your productivity. Numbers should be your main priority and the number one argument when making any business decisions.

Volodymyr’s bio:

Volodymyr Todurov is a risk manager and the head of anti-fraud and analytics at PIN-UP Global. He has 8+ years of experience in gambling and betting. His main focus is risk management — with the team, he works on developing anti-fraud solutions for sportsbooks, casinos, affiliate programs, payment solutions, and more. For Volodymyr, the DDDM approach is the number one priority — all the decisions he and his team make are based on in-depth data analysis.

 

0 / 5

Your page rank:

Plus d'articles

LAISSER UN COMMENTAIRE

S'il vous plaît entrez votre commentaire!
S'il vous plaît entrez votre nom ici

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Derniers Articles