Thought Leaders in Artificial Intelligence: Venkatesh Bala, Chief Risk Officer, Biz2Credit (Part 2)

Sramana Mitra: What is the structure of your AI model? Talk about what the architecture of the AI is.

Venkatesh Bala: We collect permission-based information when the customer comes to our marketplace. We can build a robust platform that allows us to ingest a substantial amount of data, and we can apply cutting-edge AI tools. 

Sramana Mitra: That is the most generic answer. We are a sophisticated audience. 

Venkatesh Bala: Let me elaborate. A big area that has become exciting of late is called cash flow analytics. There is a distinction in the context of small businesses. Historically, when small businesses go to a bank, the processes are manual. They would say, “Okay, fill out this form. Give us these bank statements and other information.” That is specific to a point in time.

A lot of it comes in paper statements and electronic files. They are not interchangeable, machine-readable, and scalable. Where the difference lies in cash flow analytics is that it is not transaction-level data. It’s not just numbers; it’s a lot of texts as well. In the US, for example, you have 6,000 banks and credit unions.

In addition to the ID that you have in small businesses, you also have many ID’s in terms of the information that is coming in. You can now use AI tools like national language processing and others to build scorecards and models to assess creditworthiness.

If you look at the consumer space, for example, there is the FICO score. There is no corresponding score that is universally accepted as the standard in the small business space. 

Sramana Mitra: Let me ask you some specific questions. Let’s talk about the small business lending space as it pertains to the high transaction volume category.

One of the earlier innovators of retail and e-commerce was on debt capital. They have managed to bring in data from eBay, Amazon, and other eCommerce marketplaces. That data set has allowed them to see the volume of transactions and what kind of numbers people are qualifying for.

We have seen a bunch of SaaS providers who get $15,000 MRR, and they start qualifying for cash flow financing in that category. There are players that are specifically targeting the e-commerce industry, which is a high transaction volume industry.

Where do you position your solution? We also have QuickBooks. This takes Intuit’s datasets and runs models around lending to small businesses. 

Venkatesh Bala: That is a good question. What is going to be critical for us is the tremendous amount of alternative data that is out there. You have SaaS data, accounting data, retail data, and many different things. All of these are useful in one context or another, but what we found that not all data is created equal.

The one that is closest in terms of the revenue of the business and the operational success of the business comes to cash flow. Small businesses don’t have access to capital markets. They are dependent on what comes in and what goes out. If I am at a restaurant in the northeast, the weather variable is going to be important. It’s the same in some contexts.

For example, I’ve some data on cell phone location, and it tells what is happening in terms of retail traffic. There are contexts in which those kinds of information are important. When you have cash flow data, that transcends all of these because all of the other information affects your cash flow.

Being able to understand what is happening on a granular and high-frequency level using AI tools is important. I’ve seen instances where companies would look at social media. I don’t dispute the importance of that. If I rate all those things on a spectrum or hierarchy, cash flow would be on top.

This segment is part 2 in the series : Thought Leaders in Artificial Intelligence: Venkatesh Bala, Chief Risk Officer, Biz2Credit
1 2 3 4