Artificial intelligence is a term that everybody now must be well aware of. However, not everybody realizes the vast potential artificial intelligence has and the fact that it isn’t just limited to data analytics. Also, artificial intelligence has transformed the FinTech space in various ways. In an interview with DataQuest, Rajeev Agrawal, MD, Founder and CEO, Innoviti Payment Solutions, talks about the various applications of artificial intelligence in the FinTech space, some of the trends that will dominate the BFSI sector in the days to come, and much more.
Q. How are FinTech companies like Innoviti using Artificial Intelligence/ML for business practices?
Rajeev: We are using Artificial Intelligence very heavily to improve the reliability of transactions on our systems. Different networks and banking systems behave in different manners for the same transaction. We, at Innoviti are constantly grabbing data from transactions that take place, in terms of failure rates, which bank is failing which kind of transaction, which card transaction is failing and based on this we have created our own machine learning model. Through this, we have figured out which kind of transaction, communication route and which bank has the highest probability of success.
Secondly, we have balanced this analysis with the cost that is going to be incurred in the transaction – since each bank processes transactions at different costs. On the basis of this learning, we allow our customers to balance cost versus reliability and optimize their payment scheme. With machine learning we were able to help merchants optimize their cost-reliability metrics.
As there are multiple networks which perform better in different regions, and multiple banks that have different ways of operating – we have created a model based on years and years of data – which allows us to know and counsel our customers when a transaction is happening – there is a specific route which is best suited for a successful payment. This is a model that we keep refining – hence, Innoviti has been able to achieve 99.6% reliability versus the industry average today which stands at an average of about 92%.
Q. What would you say are common challenges faced while implementing AI? Also, are you facing a skill shortage issue when it comes to AI?
Rajeev: Oh absolutely! Skillset is a huge issue. It is very difficult to get access to talented resources and, they are extremely expensive. It is definitely a big challenge; however, they are available – at Innoviti, we have been particularly very lucky and have a team of 3 people who are experts in data sciences who are able to help us facilitate the same.
Additionally, to heavily protect our ideas – Innoviti has filed for 2 patents to reassure that all we have worked on gets protected.
Q. How much of an issue is cybersecurity when it comes to automation and AI?
Rajeev: It’s a very big concern, especially when we are talking about the payment space. Fundamentally, organizations must have architectures, which do not allow any human access into the machines or data. Only machines should have access to data which has suitably been forward hashed – so that while it is unique, one cannot decode the original data out of it.
It’s like taking a field, every field needs to be forward hashed but at the same time because it is forward hashed in a manner which has a long encryption fee, it is not possible for one to decode it back at all. Data sciences gets built on top of that. Ultimately, it is not as important to know what the original value of the field was. What is critical to know is when that original value is occurring multiple times in the model. Therefore, you don’t always work on the original value, you work on a translated value.
Q. The Indian Government has recently set aside a budget for AI; would you say you are happy with it or do you feel the government should take more steps to help organizations?
Rajeev: Very frankly we do not depend on government funding today. Most private businesses are managing their business through their customers money, venture debt or venture capital. We don’t really look at that aspect. However, if you ask me, the government should ensure that the budgets set aside should be used in a combination of industry and academia together and not purely given to either one. The primary concern is about the utilization of funds and not the quantum of money. The government does set aside funds, but typically, access to private companies is very difficult and therefore, they tend to be giving it majorly to academia. Academia in India very rarely understands what the industry is currently doing. It becomes research for the sake of research and doesn’t find practical use. Balanced industry-academia participation and access to the funds can produce tremendous results.
Q. Trends in the FinTech space for this year and the years to come?
Rajeev: There are three areas according to me that will make a lot of difference:
- First, in terms of the flow of money across the supply chain getting linked
People tend to work in silos – companies that are doing B2B or B2C, lending or payments – the fundamental trend that I see forthcoming is that people are going to link how the money comes from a consumer to a business to the distributor to the supplier.
This will give birth to supply chain payments as a trend in India
- The second trend: With the birth of supply chain payments, a very significant amount of data that will get created which will have in many ways a traceability. Because you have linked the money flowing through that chain, you would automatically be “creating provenance of data”. Once you have provenance of data, it will be used for underwriting across the supply chain – is going to become a prominent trend – basically, this is the usage of Blockchain. As Blockchain fundamentally deals with that. With the growth of supply chain payments, blockchain will become more usable– it is not a very common technology that is used right now. This will lead to usability of blockchain technology coming up in a very dramatic way.
- Third trend set to happen, today the payment or money action does not influence what the customer buys, it just influences how they pay. As an example, if one goes to an outlet and buys a packet of 500gms tea – the action of what one buys and how they pay – are independent actions.
Soon brands will start using how customers pay to influence what they buy. This is specifically an interesting trend for offline retail, a tea company advertising that if a customer pays through their frequently used payment option, their MRP will always be the same.
How customers pay to influence what they buy – will be coming up in a very interesting way. This will create a new category of FinTech linked marketing platforms.