Fully automated compliance is big hurdle for this fintech's APIs
When Synapse launched in 2014, it wanted to make building a fintech as easy as building a Lego set.
Sankaet Pathak, the company’s CEO and founder, sought to develop application programming interfaces to automate every back-office banking capability.
Since then, the San Francisco fintech has been pursuing an aggressive product road map.
"We want to help people get extremely savvy about financial services, especially developers," Pathak said.
With $33 million in Series B funding, led by the venture capital firm Andreessen Horowitz, the company plans to develop a direct plug-in to Visa, an artificial intelligence-powered chatbot and a seed financing program for startups.
So far this year, the company has built a direct plug-in to Mastercard. Last year it rolled out a debit card, a white-label loan issuance product, a cryptocurrency wallet for bitcoin and other products.
The company reported 3 million end users, with 10,000 user sign-ups and 5 million API requests each day. It claims to have processed more than $2 billion in automated clearing house transactions and $40 million in card transactions this year. Synapse works with around 150 fintech customers and three banks on the back end. Synapse also plans to enter Europe and Canada with some of the products it released in the U.S.
Angela Strange, a partner at Andreessen, and Michael Hoffmeyer, director of the Crews Center for Entrepreneurship, are joining the company’s board.
Pathak spoke with American Banker about Synapse’s evolving strategy, its new suite of products and what its seed financing program for fintechs will look like. The following conversation has been edited for length and clarity.
How has your initial idea for Synapse changed over time?
SANKAET PATHAK: Our thinking has refined a whole bunch. We've done a bunch of stuff with business banking and lending. Things that I hadn't even contemplated: doing push to debit and bill pay.
I was coming out of college and, to me, something like bill pay was for a Netflix subscription or paying my rent. I've also never taken out a loan for anything, just lived with the money that I had. I've never contemplated how essential a push-to-debit product could be for people in need of emergency funds.
So we've just become a little bit more sophisticated, but the overarching goal is still the same. We just want to make it easy for anyone to sign up for a high-class financial product.
What's the biggest hurdle in creating these APIs?
Most of the people focus on building the API. But with financial services, the hardest part is back-office automation.
How do you account for every single edge case, know-your-customer, identity verification and integration into payment processors? Back office automation is the most complex part of this puzzle. The way software is built today, manifesting that with an API is easy to do, but automating the back office is the hardest thing.
What does gaining a direct processor to Mastercard and Visa help you gain?
We don't want to do what we're doing today just in the U.S.
We want to do this in pretty much any country that is not sanctioned. A big part of that is owning the process ourselves, because if we own our Visa and Mastercard processor, then we go in Europe and use the same stack. We go in Canada, we use the same stack, we go to Asia and use the same stack.
It's becoming critical to have this process and it be built directly by us, because otherwise you would have to integrate with any regional processor. In some cases regional processors don't even exist; like Asia doesn't have a good enough regional processor.
You want to make your warehouse line of credit easier to apply for. Why?
We're making it super easy for people to get up and running with the loan product right now.
As these platforms start scaling their loan programs, they have to ask an alternative lender for something like a $100 million credit line so they can loan that line to customers, and we don't want people to have to seek out a third-party lender. We just want them to come to Synapse.
What kinds of improvements are you looking to make to your ID verification and Video AUTH stack?
When it comes to computer vision, there are so many edge cases like wear and tear on the ID, lighting conditions, etc. As we are going into different regions we're adding European IDs and Canadian IDs.
On a constant basis you want to increase your optical character recognition accuracy. It's around 98% now. That's in the cases where we think a good ID is presented to us. I want to be able to accommodate poor conditions even more.
On Video AUTH, it's the same thing on computer vision. There's lighting conditions [that can affect pictures], and over time people are growing facial hair and getting glasses and are aging. On voice, it's mostly accounting for accents and making it more flexible.
It's around 98% now. Is there an industry standard for ID verification? Do regulators look for a certain level of consistency?
It's higher than 98% on IDs that are really good. It's fairly new, so the mandate is to do the best that you can. They are so limited in regulation on artificial intelligence that it's mind boggling. That needs to change at some point. Right now the regulators care that you do it to the best of your ability, avoid adverse impact and avoid fraud.
What do you have to do to get to self-service?
Some of this is about user interface, and a lot of it is about compliance. It's really building these customer identification programs and also underwriting models. The second is the user-experience piece.
This hasn't been done in our industry. This allows someone to go on our website and pick whatever products they want to build and shows them exactly how much that is going to cost with all the contracts, features and service elements right there. They accept the terms, tell us a little bit more about themselves and upload the due diligence information. It's like what Apple does to approve an app on the app store. You upload all this information, and once you’re done you just submit and publish your app after Apple reviews it and makes sure it's compliant.
As we go to different countries where we don't have an on-ground presence, this will be our distribution strategy. We won't have salespeople there. We'll have community engagement managers who will engage with the developers in those countries.
How do you want to create a seed-investing program?
It's in the very early stages, but the problem we want to solve is reducing the barrier to entry to launch financial products. Part of that is making sure you have money to build your company.
We have so much data on how well one of our customers is performing that it is easy to facilitate a seed investment with them. We have almost all the metrics you need to evaluate if this company is going to be successful or not. Then we can bring in investors who are already investors in Synapse or close friends, and share data upon this company agreeing to share data.
Plus, we'll give them a perspective on how the product is doing and where we think it will go and help these venture firms get extremely sophisticated about these companies. They would have so much more information and knowledge that we think it would heat up the investment process.