Adwait Joshi is Chief Seer and an expert on big data analytics. His firm, DataSeers, provides a big data AI appliance, FinanSeer, for banks that ingests multiple data sources and creates powerful analytics that help drive reconciliation processes, regulatory compliance monitoring, fraud detection and a full 360 view on consumer and business data.
With a passion for both technology and music, Joshi also founded SwarGanga Music Foundation, a non-profit based in Atlanta GA to promote young and upcoming artists in the field of Indian Classical Music (ICM). The organization has grown to be the largest resource for Indian Music on the entire Internet.
He holds a master’s degree in advanced data visualization from State University of New York at Buffalo, and as you can imagine, Adwait is a very interesting person to talk with about AI industries.
M.R. Rangaswami: What is the importance of Artificial Intelligence (AI) and Machine Learning (ML) in the Banking and Payments Space?
Adwait Joshi: Faster payment methods smooth commerce, but they also create the opportunity for fraud. Bad actors are continually finding new ways to breach security and steal valuable information. The modern complexity of transactions and the continual innovation happening in this space gives fraudsters new ways of committing fraud every day. As a result, rule-based monitoring simply doesn’t work. The antifraud rules have to be right every time but the fraudsters have to be right only once to get what they seek. You cannot bring a knife to a gunfight, and you cannot use rules-based engines to catch complex fraud that’s taking place using sophisticated AI algorithms.
With data growing in three dimensions (Volume, Velocity, Variety) it is very important to harness the power of machine learning and artificial intelligence in order to identify complex patterns and anomalies. Even if a faulty transaction does go through, the pattern of behavior it generates helps strengthen the information chain for the future. Eventually, instead of stopping you from doing this transaction, I can actually predict that you’re going to do a bad transaction before you do it. There’s great value in that in the long run.
M.R.: How important is the role of “data” in ML/AI over the algorithms themselves?
Adwait: It’s everything. Data is by far the most important asset banks/fintechs have. Just like how a blood test will tell you if something is wrong with your body, data can also tell you if something is out of order. But in order to be effective, the data needs to be “cleaned,” stripped of all the extemporaneous “noise” in order to identify patterns of behavior. Clean data is absolutely essential and often the most crucial step towards having an AI/ML focused strategy. This is where the data analysts come in.
Most people who work with data today are data analysts who simply clean and prep data, readying it to be used for algorithms. True data scientists, the people who look at data and figure out the elements that are important for a specific type of problem identification, are few and far between. This is because most algorithms in ML/AI are already published. Therefore, the real-time is spent cleaning, homogenizing, and labelling your data (which can then train the algorithms already in existence).
M.R.: Are Banks and Fintechs efficiently leveraging their data assets today?
Adwait: To be honest, most banks don’t know how to effectively leverage their data assets, because they lack a data-centric approach to the business. For years they have been piling up information and it has become too big or too complex to handle beyond simply looking at it on a spreadsheet. Banks are reluctant to spend money on things that don’t have a definitive ROI. Data is one such thing.
Effectively working with and leveraging data requires a fundamental shift in culture – creating a data-centric approach to the business across the enterprise, not in silos. Going into the latter half of this year and into 2020 it will be critical that banks and other fintechs have this strategy on their road map. Then, and only then will we get actionable insights that can change everything. The companies that simply “ignore” data are bound to fail.