Nethone: Fighting Against Payment Fraud With Machine Learning 2020

Nethone: Fighting Against Payment Fraud With Machine Learning 2020

I have interviewed, Hubert Rachwalski von Rejchwald CEO at Nethone, a data science Tech Startup specialised in AI-driven business intelligence and fraud prevention.

Nethone: Fighting Against Payment Fraud With Machine Learning.

Why did you create Nethone?

Nethone came into life in 2016 within Daftcode, one of the most prominent tech venture builders in Central and Eastern Europe. We have realized that there was a huge potential to equip online merchants with more data and relevant actionable here-and-know recommendations when it comes to understanding their online users. That is why we have decided to build our technological stack, which would enable us to deliver solutions related to key problems that these companies were facing. That is how Nethone Guard came to life.

Online payments market develops constantly year by year. So do fraudsters with their sophistication of cutting-edge fraudulent attacks. Therefore, we have decided to concentrate on the problem of online payment fraud and demand for less and less friction in terms of conversions and UX, which we are tackling ever since. Current reports show, that we are dealing with a huge market – in the period between 2018 and 2023, losses due to card-not-present fraud can reach up to USD 130 billion [1].

Today we are proud to have one of the most advanced systems for fraud detection and prevention, based on machine learning and our proprietary Profiler, that we are supplying to merchants all over the world. The uniqueness of the solution we are introducing bases on extraction and the consequent analyses of an exhaustive set of 3 categories of unique attributes: device characteristics, network characteristics and behavioural biometrics.

Founder personal story

I’m a graduate of the HEC Paris (Hautes Études Commerciales de Paris) and the Warsaw School of Economics. I wrote my first master’s thesis on leveraging AI for stock portfolio optimization problem. I’m a business executive with extensive professional experience earned in the world’s leading consulting firms (BCG and PwC). From 2016 I have been advising the Board of Daftcode. Since 2017 as Nethone’s CEO, I’m responsible for creating and operationalizing the company’s go-to-market strategy, coordination of key business development projects and building relationships with all stakeholders.

What is your vision?

From the beginning, our mission was to make the online space a safer, and more transparent place for all users, hence the the name Nethone – the noun “net” and the verb “to hone” which means to make something perfect or completely suitable for its purpose. That is why we decided to create a system, which guarantees exclusive transparent engagement but still preserves individual privacy. Honing the network is also about making the interactions frictionless. It’s our mission to provide the best performance with no friction.

How do you help your customers and partners?

Nethone deals with the problem of fraud in the Internet space, i.e. a situation in which an unauthorized person acquires and uses payment card or credit card data and makes online purchases on behalf of the specific cardholder. Nethone ATO is the next product, which we are developing. As it names already suggests, it will prevent the illegal takeover of online accounts (the initial focus is placed at the financial institutions).

Most often, companies accepting payments online leverage some anti-fraud features, but they are largely outdated and based on rule systems, which are reactive because rules are often created in response to attacks already made. Nethone Guard is based on access to the highest quality indicators of anomalous session, enabled in real-time thanks to the use of Machine Learning (ML) – it targets attempts of unauthorized use of credit cards and analyzes information about Internet users and flags suspicious transactions in real time.

Nethone also reacts to the changes of legal regulations, which purpose is to secure online users. Implementation of 3DS transactions authorization is such an example. Adding the next step within the user authorization process may cause a decrease in sales conversions. Nethone Guard, as a background tool, verifies the customer without contributing to the additional stages of authorization and therefore does not contribute to an increase in the percentage of abandoned shopping carts.

What makes Nethone different?

First of all, one of our most valuable differentiators is our proprietary Profiler. This is an exceptional solution, which gathers over 5000 attributes and unique fingerprints about each user and their device used to access a website or log in to a native application on a mobile device. The depth and breadth of this “here-and-now” online user profiling is quite exceptional in the branch. All these metrics improve prediction of user behavior and delivery of relevant and actionable insights in real-time. The uniqueness of our Profiler bases on analyses of 3 categories of attributes:

– Device characteristics (hardware, software, browser),

– Network characteristics (connection types, VPN detection, etc.),

– Behavioural biometrics (the way that user acts, which is completely unique for each person; the way they type, scroll, move the mouse or touchpad, etc.).

Secondly, to analyze the data gathered through the whole user flow, we use advanced Machine Learning (ML) algorithms. Attributes are analyzed by our AI models to find non-obvious links between them. ML has proven to be more effective than any other statistical rules with its proficiency to analyze highly unstructured data. The behavioral context only further enriches the accuracy of output. Moreover, we deploy a tailor-made model for each of our clients to ensure they are cohesive with their KPIs. Thanks to that, Nethone Guard provides a real-time decision on whether the transaction is fraudulent or not.

Finally, thanks to our approach Machine Learning is no longer a hard to understand another black box solution. Our goal is to explain the recommendations given by our models, through an intuitive and transparent panel where we present internal data and decompose recommendations on a set of signals, arguments and features. Our clients can (if they wish) check everything by themselves. It makes our relationship stronger, and let the merchant understands two things – decisions delivered by our model fully and why our approach is so good.

How to get in contact with Nethone?

We encourage you to visit our website and write to us via our form: or official e-mail:

If you want, we can arrange a demo presentation for you. You can contact us in this matter via our website:

Our company is still growing, therefore we have many open recruitments, which you can all check out here: or just write us an e-mail:

You can reach also via social media:

Tech Startups Magazine.

If you are interested in more tech startups, you can find in our Tech Startups Magazine and our Worldwide Directory of Tech Startups.or if you are interested to keep updated with the last Tech Startups News, write a comment below and we will contact you.


Javier Nieto León

When I started as a Business Development Manager at a Tech Startup Vendor in 2014, I faced the issues and the pains about how hard it is becoming a global Tech Startups. The Startups Tips mission is to help your Tech Startups to go from Local to Global.

Leave a Reply

KEEP UPDATED with our latest Tech Startups and Tips.