Feedzai is a Portuguese startup, dedicated to prevent fraud in electronic transactions. I had a chance to chat with their CEO & Co-founder Nuno Sebastião about the company, their plans, and artificial intelligence.
Your founding team stems from diverse backgrounds. How did you come together at Feedzai?
Feedzai is a Portuguese data science company that makes banking and e-commerce secure using real-time machine-learning to analyze large amounts of data, in order to detect and prevent fraud in electronic transactions. Feedzai was born under my MBA at the London Business School.
We have developed a software called Fraud Prevention That Learns™ that reduces the risks of online shopping and banking, either personally, online or through mobile devices, by detecting fraud through a deep historical and behavioral analysis of the organization’s data.
Every day, about $3 billion in daily transactions are analyzed by our machine-learning technology, which helps our customers – payment networks, processors, banks and retailers – to save millions of dollars in fraud losses.
Feedzai analyses well over one billion dollars in transactions every day. How did you get to this point?
Currently is more than 3 billion dollars in daily transactions. This analysis has allowed us to prevent and detect fraud in our customers’ electronic transactions, which include some of the biggest payment networks, processors, banks and retailers around the world. We come to these numbers due to several factors, but mainly because of Feedzai’s growth and the number of customer’s increase.
What are the main lessons that seven years of activity have taught you?
That we can never give up on the projects we believe in and in which we see potential. At the beginning of Feedzai’s existence, we received many denials from investors until we got to the first “yes”. We knew how to be persistent, we knew how to believe in the technology that we had to innovate the market, and it was a matter of reaching the first investors who believed in us and gave us the opportunity to grow.
How are the plans to get to unicorn status coming along?
It is true that in the startups scene the unicorn status is very valuable, as it corresponds to an evaluation that can exceed one billion dollars. However, some unicorns do not produce real products, do not have a real business or do not achieve real results. More than unicorns, I value real animals, in this case rabbits, whose name denotes, in this context, a ” Real Actual Business Building Interesting Technology (rabbit)” – I emphasize here the term “real” in the sense of “real income”. The evaluation of a unicorn is based exclusively on its potential, not on the factually generated revenue. And many technological unicorns end up not reaching their potential. It is therefore very important to know how to identify a unicorn or a rabbit, to know where to invest – time or money. Nowadays, the advertising turns out to be bigger than the reality, and as a rule it is difficult to distinguish one from the other, and in my opinion, the rabbits turn out to be more fruitful than the unicorns.
What are the challenges of managing such a quick growth?
The main challenge we face is that we have to be constantly innovating and making our Artificial Intelligence and machine-learning product at the forefront of innovation. We are aware that a company competing with a different product may emerge someday. Feedzai has grown at an amazing pace and reached the trust of the largest banks and retailers around the world, which gives us the responsibility and the challenge of presenting the most innovative technology to them and the market.
What do you envision Artificial Intelligence being capable of achieving in the future?
AI interfaces will become more conversational. As software programs that use a combination of Natural Language Processing (NLP) and Text-to-Speech (TTS) to understand a user’s voice request are developed for chat bots and intelligent assistants, the way we interface with AI becomes more conversational, with both textual and especially audio means becoming central to the way we get information from AI systems and input data in general. As a result, AI goes from blackbox to “whitebox” processing where machine logic is demystified via human-understandable language. Could this signal the death of the keyboard?
Not quite yet. For the near term, we will still be typing and talking at least. Using web searches as a benchmark, compared to a typical text-based web search, it takes 165 times more computing processing resources to perform a speech-based search (converting speech to text, performing the search, converting results back to speech).[1] Fortunately, faster processors, in particular FPGAs and GPUs used in for machine learning and other AI processing methods, are becoming more widely available, with the market expected to grow at a CAGR of 9.1% from 2014 to 2020.[2]
Also, machine learning and AI platforms, emphasis on “platforms,” is what everyone is betting for 2017. Platforms are perceived as much more complete solutions and you can charge a premium for solutions that are “platform-ized”. The cost of point solutions will be difficult to manage and although most companies will buy solutions for analytics, Big Data and machine learning model development, these strategies will quickly realize that the lack of portability and integration will cost organizations in their ability to respond to market conditions and adapt quickly.
Where do you see the Portuguese startup scene in five years?
The Portuguese startup scene has gained a greater prominence in recent years, with the emergence of new and innovative technology-based companies, achieving international repercussion and recognition. Portugal counts on with very talented people, engineers that are demanded by the biggest world companies. Entrepreneurship has been boosted at several stages and the coming of Web Summit to Lisbon is also a driving force putting Portugal on the radar of Entrepreneurship on a global scale. I hope that five years from now we will continue to have stories of successful Portuguese startups worldwide and that there will be more benefits and incentives to entrepreneurship in Portugal.
If you weren’t working on Feedzai, what would you be working on?
When I’m not working at Feedzai, I enjoy hand building audio electronics.
[1] https://blog.acolyer.org/2015/04/13/sirius-an-open-end-to-end-voice-and-vision-personal-assistant-and-its-implications-for-future-warehouse-scale-computers/
[2] http://www.grandviewresearch.com/industry-analysis/fpga-market
Leave a Comment