My co-founder and I launched similar products 7 years apart, here’s what we’ve learned

On trying again

An ancient Greek story tells of Bellerophon’s attempt to tame Pegasus, the legendary winged horse. He first tries to capture Pegasus by ambush at a watering hole, but Pegasus easily escapes.

Afterward, Bellerophon rests at Athena’s temple, where she appears to him in a vision and gives him a golden bridle, teaching him how Pegasus must be approached. Following her counsel, Bellerophon uses the bridle and successfully tames Pegasus.

No golden bridle

Ten years ago, my co-founder and I stood on the stage at Finovate Fall in New York to announce Unbill. Shortly thereafter, we announced the payment switching technology CardSwap at the same event. The product did fine, but didn’t really take off. Still, many people believed there was something there — we just hadn’t approached it the right way. When Atomic went live with the largest neobank in the U.S. in August of 2020, Atomic performed what we believe to be the first automated direct deposit switch in production. We refined and improved that product over the next few years, but many of our customers asked us, “Are you going to take another pass at payment switching?” Our answer in the early days of Atomic was, “We don’t think we could do it better or differently than last time, so no.” We had no golden bridle—no real learnings on how to approach things differently.

Discovering our golden bridle

In 2021, we began working on what would become our TrueAuth technology. We announced this technology in February of 2022. TrueAuth allowed us to approach our connections to payroll systems with improved security, something we felt would be critical for the most important clients we were selling to. What we didn’t expect was the dramatic lift in conversion rate. In September of 2022, we shared that “users are +32% more likely to authenticate successfully via Uplink when compared to traditional approaches.” In the years following the development of the TrueAuth, our company went on to secure relationships with hundreds of FinTechs and financial institutions — including 9 of the top 10 banks — largely due to this significant technological leap.

But people kept asking, “Are you going to take another pass at payment switching?” Our answer began to change in 2022. The TrueAuth technology works the same at ADP and Amazon, Ceridian and Crunchyroll, Paycom and Peacock. So we started experimenting. We found that this new way of authenticating into these systems yielded significant security gains compared to traditional screen scraping methods and led to much higher conversion rates as well. For example, users could now leverage their password managers when permissioning access to merchants, something our prior technology couldn’t support.

Today’s agentic systems often struggle most with authentication. There are captcha’s, tests of human intelligence that make us wonder how far off we are from the average IQ, and a slew of other barriers block systems from accessing user accounts. We discovered that the future of acting on a user's behalf starts client-side — on the device. LLMs have encountered similar problems and responded by creating their own browsers. But much of what our customers want to accomplish is within their own apps or environment.

After some rest and approaching the problem with this new technology, we felt we’d found our golden bridle.

We began leaning fully into the TrueAuth technology, earning two patents for this innovation. Along the way, we discovered far more use cases than we’d originally considered — largely due to the client-side nature of the technology. Sessions initiated by a user and kept on their device were long-lived and came with significantly improved cost structures compared to holding sessions open on AWS servers. This gave us the ability to check for savings opportunities, sync purchase data with major merchants like Amazon, Walmart, Target, and Costco, and much more.

In September of 2023, Atomic announced the launch of our payment switching product — seven years after announcing Unbill at Finovate Fall in New York. We did so because we believed we’d finally found our golden bridle: a new approach to merchant connectivity that could be truly game-changing for the industry. Atomic is approaching merchant connectivity — inclusive of payment switching — in a vastly different way, and it finally feels like we’re taming Pegasus — the market itself.

Why now?

While companies like Yodlee, Finicity, MX and Plaid paved the way for user-permissioned access, Atomic paved the way for user-permissioned actions, what many people today refer to as agentic workflows. We’ve seen the largest fintechs and banks in the world adopt and come to rely on automated direct deposit switching. Today, automated direct deposit switching has become a mainstream capability that consumer-focused financial institutions must have to compete and, more importantly, it consistently boosts deposits!

After helping the industry write the playbook on user-permissioned actions via direct deposit switching, we saw the opportunity to do the same for helping consumers manage the payment method on-file with their merchants, subscriptions, and even automating payments themselves. Our team was uniquely positioned to tackle this new opportunity with the TrueAuth technology.

Riding Pegasus

With our golden bridle and new approach, we feel a bit like Bellerophon after taming Pegasus. He immediately went on to defeat the Chimera and win battles in Lycia. The feeling of catching the updraft in a market is incredible. Over the last year, we’ve worked with several of the top 10 banks to launch this product, along with many of the largest fintech companies.

Lesson learned

I’m a horseback rider. Maybe that’s why the story of Pegasus came to mind. Some would have balked at doing a product again that didn’t quite hit its stride the first time. I’ve learned that when you get bucked off a horse, you get right back on — because if the horse realizes that all they have to do to catch a break is buck you off, you reinforce the behavior.

Just because it didn’t work perfectly the first time doesn’t mean a second pass won’t work better, especially if you have reason to believe your approach is vastly superior than the first time around. We’ve had more traction in the first two years working on payment switching this time than the six years spent working on it the first time around!