Editorial Team

The Braintrust: Gil Akos (Astra) on Automation, Frictionless Transfers, and Fintech's Next Stage

For generations, people had a relationship with one big bank. Often, the bank you chose was the one in town — people’s rationales for why they banked where they did was often as simple as: “It’s because my parents and their parents banked here too.” The incumbent banks have relied on this stasis; it’s allowed them to create accounts that act more like a shoebox than a wealth-building tool.

But the last few years of Fintech have seen a flood of new challenger banks and Fintech applications that are changing the way people interact with their money. It means the average consumer has relationships with five relationships with financial institutions these days. Suddenly, it’s become imperative to be able to move your money seamlessly between accounts and Fintech services.

That’s where Astra comes in. The startup founded by Gil Akos and Sam Morgan delivers an API to banks and Fintechs that allows automated transfers between accounts, whether it’s me-to-me, B2B, or anything in between. On top of the payment rails, Astra also has built a robust automation engine, which allows consumers to set parameters for how and when they want their money to move: anything over $10,000 in your primary account can automatically be sent to your investment account at the end of the month, for example.

Unifimoney has partnered with Astra to allow our users to move money between accounts securely and without hassle. We believe automation can make wealth building effortless. We believe Astra’s API is an integral part of making investing and money management, simplified, unified and diversified. 

We spoke with Gil Akos to better understand how automation and the removal of friction in money transfers will help spark the next stage of the Fintech revolution. 

Let’s start at the start. What led you and Sam to the founding of Astra?

My background is not at all in the Fin or Tech space directly — I actually have a background in architecture. The joke is: atoms, not bits. But I ended up doing a post grad at Columbia focused entirely on computational geometry and complex systems, so I was moving more and more into technology and computing. After that, I started a business wherein we needed the ability to move money between accounts, which were individual accounts or business accounts of the partners in the LLC, and we needed software that did that somewhat programmatically over time. In thinking through things as systems, the persistent question became “how can you route funds between accounts over time?” I was building prototypical software to do that, but just using it for my small business at the time. 

Around then, I met my cofounder, Sam, who has a background in economics and finance — he likes to joke that he was the automation at Deutsche and Morgan Stanley in his early professional years. So, together we created Astra based upon the notion that there should be software for individuals and businesses that moves funds to the right place at the right time, and allows you to do so with some degree of rules and automation. Software should exist that takes the cognitive load off your plate by allowing the system to operate across your entire collection of accounts, moving funds back and forth automatically and for maximum optimization.

There’s a tendency towards stasis that incumbent banks rely on. How can Astra help to remove that friction that keeps many people at banks that deliver less value for money?

Yeah, the way we think about it is that we try to map and categorize each end user's financial graph: the nodes being accounts that they have at various institutions and edges being the flow of funds between those nodes. We've seen that people will frequently sign up for new accounts, but they don't necessarily wind them down. That just means that we have more accounts to manage over time. Because of that, we need some layer of functionality that operates across them all. And that's the need that Astra is trying to solve for.

A really interesting data point I've seen is that in 2015, the average consumer had, on average, two relationships with financial institutions. Today, that's already grown to five, and it's expanding rather quickly. All the amazing challenger banks like Unifimoney and new Fintech products out there are a key part of that trend.

I spoke with Alex Johnson who writes the Fintech Takes newsletter who explained the concept of automation in money by using the framing of automation in cars. We're at a place in car automation where you probably could have a fully self-driving car, but consumers are more comfortable with a self-driving car where you still sit at the wheel. Within that metaphor, where does Astra see itself on the automation chart? Can you guys move the wheel when the car starts to swerve? Or does the driver not need to be there at all?


Money and finance is abstract, and the idea of automating money adds another layer of abstraction. So, frequently, we look to analogies to help explain exactly what it is that we do. And I think the self-driving car one is spot on. 

The one thing not spoken about enough when you're talking about self-driving cars is the sensing and mapping technology that helps power the automation. There are six levels of driving automation — the ultimate is full automation, where you don't have to interact at all. With Astra's current capabilities, we are definitely at level three, which is conditional automation, and moving into level four, high automation. This is the bleeding edge of automation within finance and there are a few factors that let us get there. 

First, you have to have a broad map and the ability to adapt your route. It's Google Maps paired with Tesla's automation that really makes it possible to get to higher automation or beyond. While it's totally okay for us to have an autonomous vehicle in a closed loop system, like a parking garage or a planned community where there are bounded edges, when you're on a highway, where the conditions are dynamically changing — for example, an accident ahead that wasn't there three days before — there has to be the ability to reroute and manage the change in context in real time. So, you have to have both the map and the capability to reroute and manage things when there are edge cases that come about. We've spent two years building technology to gracefully handle edge cases, and expanding the map and the routes that you can automate money across. 

It's really exciting to see how we get to Level Four and Five, though we know that in a lot of cases — especially consumer cases — we don't really want or need to go past Four. But Level Three or Four automation means that Astra users no longer have to be 100% focused on the next 100 yards of road at all times. Anyone on a shoestring budget knows that level of vigilance is stressful. With Astra, you can kind of zone out for a minute and let the automation take the wheel. When you do happen to swerve out of the lane, when there's an edge case that requires action, you're prompted to take that action. That's a really good fit for the vast majority of use cases in finance.

There’s a story about Betty Crocker: they tried to sell their cake mixes to which you just added water and it failed. A marketing guy proposed the idea of removing one dehydrated egg from the mix so that home bakers could feel like they were actively baking by cracking an egg. That’s what launched the runaway success of the product. Consumers want automation and want ease of use, but it seems they also want to feel their hands in the batter. How do you guys balance that at Astra? 

I love that analogy! I hadn’t heard of that one before. My answer would be twofold. On the technology side, the ultimate opportunity for us is to be a platform to enable the widest set of use cases. So, we are working systematically to expand a set of use cases, whether it's me-to-me as the transaction route, peer-to-peer as the transfer route, B2B as a payment solution, or anything else. But like AWS, there will also be lots of different things you can do that people didn't perceive were the right use cases for the underlying modular technology elements upfront. From a consumer interaction point of view, one application we’ve thought about it is that eventually, perhaps the user signs into the Unifi app and Unifi says, you should do this, this and this, click a button to approve — that could be interesting. 

But I think an important touch point for the end user is that they always specify, in some way, the parameters that describe the automation itself. Though you're not the one putting a module of business logic into Astra's automation engine, as a consumer, you are the one mapping your intent to what should happen when. You're the one saying: when my account balance at Chase goes above $10,000, at the end of the month, move those funds over to my Unifi savings account. It's you who's setting the brackets and parameters around what should happen. That transfer will then happen through our automation engine, but I do think that's a really key part. 

What we're in the business of doing is making the Betty Crocker mix, but we let the end user crack the egg. They enter, in a guided fashion, what those variables should be and bundle that into an API request that then gets logged into our automation engine which turns into transfers on a user's behalf.

The first decade of Fintech has been about solutions on the margins of finance. How can the rise of viable neobanks and the ability to more seamlessly switch change the way the average consumer interacts with their money?

I think that's a really astute observation in terms of where we are. If you go back 15 years to the rise of Fintech 1.0, it was about new access points to data. Right? So, the first generation of products, like Mint, were built around that new data. The second wave, Fintech 2.0, was about access to new infrastructure — think: Plaid and Stripe. The ability to securely connect into the data points for an end users' bank accounts by way of Plaid was huge. 

That led to this next stage, based on the, in retrospect, obvious outcome: if you have access to this infrastructure, you can build affinity-based products, either centered around functionality or community or demographic. Because of the access to the APIs and platforms, the capital expenditure is no longer so high that it precludes new entries into the market. So, FinTech 3.0 is going to be defined entirely by access to platforms and specifically not modular API's for X, which are important and the key outcome of FinTech 2.0. That means there is now the possibility for a Cambrian explosion of product types, categories, and use cases.

Astra’s position as a B2B solution gives you a fascinating view of the entire banking and Fintech ecosystem. Where is the next game-changing innovation coming in personal finance?

I will qualify that I'm biased, but I think if you take a parallel to web commerce — which is a few years more advanced than Fintech — companies like Shopify have done an amazing job of reducing friction across all the Shopify-powered stores. If I bought an Oakland A’s t-shirt at Store X, and I now want to buy a Cleveland Browns baseball hat at Store Y, as long as they're both powered by Shopify, I get the benefit of my persistent profile. With one or two clicks, I can check out, because I have payment information and shipping addresses stored from previous activity. It's better for the web store, because it means conversion from card to checkout is a more efficient process. And it's good for me, because it's really easy for me to complete the purchase. Obviously, it's also good for Shopify, because they're facilitating the sale. 

I think that's a model that is very recently emerging in Fintech. Astra is absolutely a part of that trend as well. You can single sign on and you've authorized your account over here with Chase, and you're doing so with Unifimoney, and you could do so again on the next application so long as it's also powered by Astra. That reduction of friction is a net benefit for everybody. That's the true definition of a platform and why I described FinTech 3.0 as being defined and probably mostly won by those that can really develop and build these platforms that enable what other developers can build!

JOIN

Join our newsletter

We are now slowly rolling out our beta program. Be one of the first to get access by signing up today.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

*Important information and disclaimers

The above does NOT constitute an offer, solicitation of an offer, nor advice to buy or sell specific securities. The opinions listed above are not the opinions of Unifimoney Inc. or Unifimoney RIA, Inc. but represent the opinions of independent contributors. These contributors may or may not hold positions in the stocks discussed. Investors should always independently research any stocks listed and form their own opinions, while recognizing that any investments made may lose value, are not bank guaranteed and are not FDIC insured.