Assist Antifraud

2015: Assist Antifraud

Assist Antifraud is an intelligent system of counteraction to fraud for use within online store.

The principle of formation of mathematical models of assessment of probability of fraud is the basis for the product Assist Antifraud. A payment gateway, having data on "net" and fraudulent transactions for a certain time frame, can use them for setup of mathematical models. Models for fractions of a second compare the entered transaction with all volume of earlier made transactions and determine the probability with which it is similar to fraudulent, at the same time the analysis of thousands of parameters characterizing the client's profile, his behavior on the website, payment history and data of the specific order is carried out.

On December 15, 2015 the Assist company announced upgrade of a system of an antifraud and increase by means of a product of conversion of online store up to 99%.

According to the statement of Assist Group, the efficiency of an intelligent system of counteraction to fraud is increased that gave the chance to bring conversion of online store to 99%. At the same time the level of a fraud is no more than 0.035% without use of filters and 3D-Secure.

Scheme of interaction of the Assist Antifraud system

Scheme of system operation

1. Scanning of a profile of the client (collecting of the maximum quantity of the data characterizing a customer behavior at implementation of purchases and payments).
2. Data validation, creation of communications between them (formation of a numerical profile of the client).
3. Mathematical analysis of a profile of the client (determination of statistical probability of fraud).
4. Depending on an analysis result the payment can be rejected as suspicious, is sent for processing in bank or will subject to additional check.

In composition of Assist Antifraud the subsystem of business rules and a mathematical definition of probability of fraud works, it is equipped a set of the reports helping specialists of groups to analyze a fraud monitoring suspicious transactions. In 2015 operation on creation of an independent OLAP system which puts all data of the reports used earlier is carried out. It helps to create the multidimensional cuts of data, necessary, both for the analysis of transactions, and for identification of fraud on the party of online stores. It helped to increase high-speed performance of a system and the number of the created selections.

The question of security of payments on the Internet cannot be considered separately from its influence on sales volumes. Business needs such method of protection which at admissible risk will reduce minimum sales. The efficiency of a system directly depends on the saved-up history of transactions and experience of specialists in an antifraud who are very difficult for finding in the market. Experience is a core competency because formation of rules of functioning of a system is performed by forces of experts. In ASSIST specialists work with more than 15-year experience with card transactions that is the most valuable company asset. Vladimir Anisimov, director of research and software development of Assist group

The main quality of the antifrodovy systems created on the basis of mathematical models – a possibility of adjustment under specific objectives and type of business. Training of model at the history of transactions of that online store for which it is intended is for this purpose provided.

Use of intelligent systems of counteraction to fraud does not exclude application of a set of means of an antifraud (3D-Secure, filters, "manual" monitoring), but optimizes it, helping to prevent ~ 99.97% of fraudulent transactions.

2016: Assist Data Science OLAP

The provider of electronic payments Assist announced in 2016 a multifunction OLAP system which is used as for protection of electronic payments against fraud, and for creation of marketing tools for increase in efficiency of sales. A system is developed by means of the Data Science methods and allows to solve within one interface a number of key problems for the enterprises of e-commerce – to ensure financial safety and to increase turnover.

An OLAP system provides operational functioning of the intelligent Assist antifrodovy system constructed on use of mathematical models. A system analyzes a set of the parameters characterizing the client's profile, including, his consumer activity, payment history and data of orders. Assist as a payment gateway, more than 15 years performing carrying out electronic payments, has an array of the saved-up data on fraudulent and "net" transactions for all history of a company performance. On the basis of these data mathematical models which allow to define legality of transaction with efficiency to 99.97% without use of filters and 3D-Secure are under construction. Creation of the models and report execution which are built in the system of order fulfillment on the website is performed less than in 1 second.

Also an OLAP system is used as the instrument of obtaining multidimensional data in the marketing purposes. For example, on demand conversion and analytical reports are provided to the companies. The conversion reporting contains data on the number of successful transactions and also about payment reasons for rejection. Analytical reports form on the different cuts allowing to estimate sales volumes and their dynamics for the selected periods, distribution of buyers by a floor, by devices from which orders, etc. are performed. A system provides a convenient service on obtaining such reports – the order is performed in a personal account, the report forms quickly and sent by e-mail.

By means of the Assist OLAP system creation of the reports allowing to analyze the payer's profile for the purpose of increase in efficiency of sales of online store is possible. For example, such parameters as are investigated: the amount of purchases of the client for a certain period on one or several shops, the frequency and an average bill of purchases, commodity categories, the preferred payment methods, types of payment cards and their issuers and so forth. Using such reports it is possible to study set of clients of online store, to segment it by different criteria, to analyze traffic deviations from regulation, to estimate solvency of buyers.

Assist uses an OLAP system applicable to all transactions which are carried out through the payment gateway, ensuring safety of payments for more than 3000 companies of Russia and the CIS among which there are Google, OZON, OZON.Travel, Mariinsky and Big theaters, KASSIR, Bileter, Head Hunter, PUMA, OTTO, Rosgosstrakh, The Body Shop, Mothercare, KFC, Mosenergo, Softline, Allsoft and many others.