The fintechs solving B2B commerce
In COLLABORATION with
The fintech revolution may have come to your (digital) wallet—but some businesses are still stuck in the past. A little over one in four B2B payments in the US and Canada are still done by paper check, for instance; and when it comes to procurement, financing, and regulation, businesses constantly run into manual processes that haven’t changed in decades.1 Things are starting to shift, however, thanks to a wave of fintech companies with smart solutions to age-old problems. We spoke to four of them about the challenges they’re tackling.
The problem:
How much can a business trust a new customer? For suppliers, who often deliver goods and services upfront, it can be an existential question. They previously dealt with it by bombarding potential customers with paper forms designed to assess credit risk—the start of an opaque “Know-Your-Business" process that could last weeks and was prone to human error and fraud. “It was a massive challenge for internal credit departments to construct a form that’s manageable and consumable by clients, and that shifts with the changing requirements that each of the countries you sell in might have,” says Brandon Spear, CEO of TreviPay. “It just became a pretty burdensome thing.”
The solution:
TreviPay is designed to lower that burden. Buyers see an intelligent digital application that adapts depending on their answers and location—and that autofills from publicly available data to speed up the process. TreviPay then uses that information to offer the buyer an appropriate amount of credit to use with a given supplier almost instantly. “These applications happen when someone wants to make a purchase, and they really don’t want to wait until next week until they have a line of credit to be able to make it,” says Spear. “If that happens, they might go to a competitor, and you might lose the sale.” TreviPay’s real superpower is its decades of data and AI-powered decisioning, which helps it assess risk based on experience.
Brandon Spear
CEO, TreviPay
What’s been the hardest thing to get right while building this?
“There’s an arms race with bad actors and their use of AI. We’ve used machine learning models and AI tools for a long time to detect fraud, but AI has really allowed bad actors to get much more sophisticated. Two years ago, they might have created a fake website that was one or two layers deep—now they have one that’s six or seven layers deep. It looks real. So, you really have to validate.”
The problem:
There are five building blocks in the procure-to-pay (P2P) cycle, the end-to-end process by which companies buy goods and services: a requisition, a purchase order, a goods receipt, an invoice, and a payment. “Initially it was all paper-driven,” says Rajiv Ramachandran, Chief Product Officer, Invoice to Pay, at Coupa. “And when it was originally automated, each block was put in different siloed systems: ordering in one, invoicing in a different place, payments on its own.” That’s better than paper, but still bad for transparency—meaning it’s hard for companies to get a handle on what they are spending across the business as a whole.
The solution:
Coupa offers a single platform that caters to every step of the P2P process, helping companies analyze their spending and optimize it. That might mean taking advantage of bulk discounts if multiple departments previously had separate deals with the same supplier, for instance, or obtaining early payment discounts by being able to process things faster. “There are patterns that only become visible to our customers when they see the data in front of them,” says Ramachandran. Coupa also leverages the data from the trillions of dollars of transactions that have run through its platform. It uses this to build pattern recognition tools that scan for duplicate invoices or potentially fraudulent payments and can identify potential cost savings.
How will this conversation be different in ten years?
“You’re going to see agentic AI transform this process, making it fully autonomous and giving people not just the knowledge they need to do more, but the time they need to do more. All the inefficient steps we talk about today will go away.”
The problem:
Imagine you’re a business selling Christmas ornaments online. Most of your sales happen in November and December, but you need to pay for your stock, distribution, and marketing long before you have any money coming in to cover it. In the past, a company of that size could be offered a working capital loan, but over the last two decades, that same ornament distributor may have had a harder time securing the capital they need, according to Aidan Corbett, CEO and Co-Founder of Wayflyer. This is because the cost of underwriting—evaluating the risk of extending credit—often outweighs the potential return.
The solution:
Wayflyer is a small business lender that automates the underwriting process. It taps into data that these businesses already have—sales volume, customer loyalty, return rates—to assess lending risk. It also pulls in data from other sources, including online reviews (falling scores are often an early sign that a business might struggle to pay its debts). Wayflyer’s underwriting model uses that information to quickly decide whether to lend and to adjust the pricing accordingly. “For lots of businesses, if we can say, ‘We can give you a yes on $250K in the next two hours,’ that feels like magic compared to what they’ve experienced historically,” Corbett says.
Aidan Corbett
Co-Founder and CEO, Wayflyer
What do people get wrong about this space?
“I don’t think people understand that the vast majority of small businesses will have to take a loan. It’s not a sign of weakness; it’s actually the nature of the game. It’s impossible to scale without outside money because you have to spend so much upfront before you generate revenue.”
The problem:
“B2B payments are incredibly messy,” says Dean M. Leavitt, Founder and CEO of Boost Payment Solutions. Multiple payment methods, different terms for different suppliers, numerous approval steps, and complex bundling are common obstacles. Take a single payment of $500 million to a parts manufacturer. “That one payment might house 5,000 invoices, and each invoice might house hundreds of line items.” For suppliers, properly accounting for those payments means manually matching each of those hundreds of line items to an individual charge.
The solution:
“We take what’s a very complex, very manual, and frankly very error-prone human process and convert it into more of an automated, smoother process,” says Leavitt. He describes Boost Payment Solutions’ core product as a commercial card funnel: At the top, it ingests hundreds of payment formats and the data that suppliers need to get paid. At the bottom, it delivers automated reconciliation that matches incoming payments to the inventory shipped. This cleans up the mess for both buyers and suppliers.
What’s the most unexpected thing you learned while building this?
“Inertia is an incredibly powerful force. It’s a lot harder than you would think to implement change or get an organization to consider implementing change. But if you can make the process easy and demonstrate real value, you can overcome that inertia and drive impact.”
https://www.jpmorgan.com/payments/payments-unbound/sources
Illustrations: Doug John Miller
