Prashant Fuloria: CEO at Fundbox Explains how Fintech Platforms Can Leverage Machine Learning to “Democratize” Access to Funding

 

We recently connected with Prashant Fuloria, CEO at Fundbox,  an AI-driven credit platform for B2B commerce.

Fundbox focuses on “disrupting” the $21 trillion B2B commerce market by developing the first B2B payment and credit network. Sellers (of all sizes) are able to use Fundbox’s platform to quickly increase average order volumes (AOV) and improve close rates by providing competitive net terms and payment plans to their small business clients.

Fundbox’s management notes that the company has made substantial investments in machine learning and the ability to quickly analyze the transactional data of small companies. Fundbox is “reimagining” B2B payments and credit products in “new category-defining ways.” Our conversation with Prashant is shared below.

Crowdfund Insider: What are the main products and services that you offer?

Prashant Fuloria: Fundbox is a financial services platform powering the small business economy with innovative credit and payment solutions.  We work with a broad spectrum of industries and segments, providing small businesses with access to working capital and other tools designed to give them greater financial agility and peace of mind.

We’re focused on serving small businesses that are on the smaller end of SMB scale. So businesses ranging from a sole proprietor all the way to businesses with up to a hundred employees. Our customers’ annual revenues range all the way from $50,000 on the low end to $10 million on the high end, and the average is roughly $750,000.

At Fundbox, we’re focused on re-imagining credit and payments for small businesses using data and machine learning. One of the biggest concerns for small business owners is maintaining steady control over their cash flow.  And while cash flow management is the core pain point, lack of cash flow control bleeds over into other areas of concern such as the ability to make payroll, buy inventory, pay taxes, accumulate savings, and more.

We want to continue building innovative products that put our small business customers in a position of strength by making them more financially agile and resilient. While we cannot provide details at this stage, we are indeed working on new products that address financial challenges small businesses face on a daily basis.

Crowdfund Insider: Do you intend to introduce any new products? What is your team working on these days?

Prashant Fuloria: At Fundbox, we’re focused on re-imagining credit and payments for small businesses using data and machine learning.  One of the biggest concerns for small business owners is maintaining steady control over their cash flow.

And while cash flow management is the core pain point, lack of cash flow control bleeds over into other areas of concern such as the ability to make payroll, buy inventory, pay taxes, accumulate savings, and more. We want to continue building innovative products that put our small business customers in a position of strength by making them more financially agile and resilient.

While we cannot provide details at this stage, we are indeed working on new products that address financial challenges small businesses face on a daily basis.

Crowdfund Insider: During the ongoing COVID-19 pandemic, many lenders stopped issuing new loans and some even shut down their operations permanently.

However, your company kept lending and you claim that you’ve managed to outperform the overall market. Please explain how this was possible.

Prashant Fuloria: The pandemic has created challenges for small businesses and all companies that serve them, including fintechs. Many small business credit providers have suffered from increasing delinquencies, the percentage of payments you expected that you didn’t get at any point in time.

We’ve seen from publicly available filings that, in some cases, those delinquency rates have gone to double digit percentages with some as high as 30% or 40% or more.  When this happens to a credit provider, it causes all sorts of business and operational challenges. Most credit providers originate loans using other people’s money—credit from other facilities—and there are very strict contracts with debt covenants tied to delinquencies and defaults.

Debt covenants are agreements between a company and a creditor usually stating limits or thresholds for certain financial ratios that the company may not breach. In the case of those fintechs providing credit to small businesses, many saw their delinquencies rise. This, in turn, caused their debt covenants to be breached, and their credit facilities getting pulled or frozen.

Which meant that they were no longer able to originate credit, essentially cutting their customers off at a time when they needed credit the most.

By contrast, at Fundbox, we only saw a modest uptick in those delinquency and default metrics in the early weeks of COVID, and those metrics quickly returned to low single-digit percentages – at pre-COVID levels and lower.

Our superior performance has been driven by three factors, the most important of which is our investment in data. First and foremost, we’ve spent over a hundred million dollars over the last 7 years to build out our data assets.

This includes business transaction data that we gather from our customers, which populates our business graph – Fundbox’s representation of businesses in the US and their interactions. This also includes the time and effort to analyze our data and generate meaningful signals (or features, as they are called in machine learning). Not only does this investment let us provide credit decisions in less than a minute (a great customer convenience), it also ensures that we make incredibly robust predictions about the health of a business customer.

Second, we’ve also invested heavily in our engineering and product efforts to fully automate our underwriting and risk management. This lets us manage Fundbox’s risk exposure effectively without slow and error-prone manual intervention.

Third, our approach to credit management and customer acquisition has created a diversified customer base. We are not overexposed to businesses in any one particular geography or vertical, a strategy that results in a more robust customer base and credit portfolio.

Crowdfund Insider: Fundbox has invested $100 million (over the course of 7 years) in data modeling, which you claim helped avoid the worse economic effects of the Coronavirus crisis.

Explain how data modeling has helped you operate more effectively.

Prashant Fuloria: Over the last seven years at Fundbox, we’ve invested over a hundred million dollars in our data asset. This investment is made up of three things. First, we have a large team of very talented data scientists, machine learning engineers, and data operations people. Second, we have invested in data-related software and services, as well as purchasing data from third-party sources. Third, we have trained our machine learning models with thousands of real-life defaults, each of which has a financial cost.

The third point may require some explanation. Right from the very start of Fundbox, we decided to make credit decisions based on predictions from statistical models, and not traditional underwriting heuristics. That decision came with a price – we incurred losses on defaults that we could have avoided by “tightening our credit box” using old-school rules.

However, the upside of our approach is that we now have machine learning models that are robust across a broad spectrum of customers, as characterized by revenue, years in business, vertical, business owner FICO score, and so on.

Let’s take a simplified example to bring home the point even more explicitly. An underwriter might use a heuristic such as “do not approve any business that is less than 3 years old”. That may reduce losses, but it also results in not serving an entire category of customers – newer businesses that are looking for capital to grow.

Our strength in machine learning, coupled with our access to real-time data ensured that we were able to quickly revise our risk assessments on small business customers as the pandemic unfolded. This let us continue originating working capital to our customers in a responsible way.

Crowdfund Insider: Please explain how your platform leverages AI and machine learning to make decisions about providing finance options.

How do these new technologies help the lending sector?

Prashant Fuloria: When a small business works with Fundbox, it connects a “transactional system” that we use as a real-time data source. This “transactional system” can be its accounting software, their invoicing system, or even its bank account.

Fundbox immediately pulls business performance from this data source and our machine learning algorithms quickly assess business risk so that we can make a credit decision. The entire process typically takes a few minutes or less.

This is an extremely easy process for the customer because we do all the heavy lifting on our end. We have automated the entire risk-assessment process, without the need for human intervention. Our systems can analyze millions of financial transactions such as deposits, payment of invoices, payroll, leases, equipment or inventory purchases, etc. – faster and more accurately than human underwriters could at scale.

And, for those businesses that are approved, we can provide access to a revolving line of credit up to $150,000 as soon as the next business day. A big factor of our ability to quickly analyze data and then provide a credit decision in minutes is our proprietary data asset called the Small Business Graph.

Much like the Facebook graph that provides valuable relationship insights among people or the LinkedIn graph provides insights among professionals, the Small Business Graph maps millions and millions of transaction-based interactions between a small business and the other businesses it transacts with on a regular basis.

This unique data asset provides us with the ability to look at a small business in real-time and within the context of its network, instead of relying on static pieces of information. And, that’s why we’re able to make very good decisions related to which businesses we should extend credit to and how much.

Crowdfund Insider: How important are small businesses to the US economy?

Prashant Fuloria: According to the SBA, 99% of all businesses in the U.S. are small businesses, employing 59.9 million people which is roughly 47% of the private sector workforce.

As per the Office of Advocacy at the SBA, 44% of the U.S. GDP is driven by SMBs. Small businesses are indeed the backbone of the U.S. economy. And given the impact of COVID, now more than ever, we need to find ways to support the health and wellbeing of our small business communities.

Crowdfund Insider: How vital are small businesses to the economies of both developing and developed nations?

Prashant Fuloria: Since we only serve customers in the United States presently, we’re less familiar with other markets. That said, both macroeconomic data and our anecdotal observations suggest that small business owners around the world share similar challenges and concerns as their U.S. counterparts: cash flow is their lifeblood, and is also usually what keeps them up at night.

Crowdfund Insider: What can Fintech lending platforms do to assist small companies so that they can become key contributors to an economy?

Prashant Fuloria: Fintech credit platforms have a unique opportunity to “democratize access to funding” and to serve those small businesses that have remained underserved.

A key reason why small businesses have been so underserved is because most banks can’t profitably serve them through traditional, manually-intensive processes that can also be prone to error and bias. This is especially true if a small business is seeking a relatively small amount of funding.

Fintechs, by contrast, can use technology and automation to make serving customers faster and less expensive. In particular, fintech platforms have a real opportunity to transform the small business economy through faster access to credit and real-time payment flows.

By leveraging data and technologies like machine learning, fintechs can break down barriers and remove friction from financial transactions so businesses of all shapes and sizes can reach their full potential. And that’s pretty cool.

Sponsored Links by DQ Promote


 

You may also like...

Send this to a friend