Core technology remains a huge challenge for the banking industry, with a majority of financial institutions still running on mainframe-based systems because switching to other options is such a monumental task. In this episode of Talking Banking Matters, McKinsey payments sector expert Roshan Varadarajan speaks with one of the entrepreneurial technologists who seek to solve this challenge, Zeta cofounder and CEO Bhavin Turakhia. Zeta offers a broad array of banking technology, with a view to offering a unified stack that powers a financial institution’s products on a single platform. The following edited transcript shares highlights from the conversation. For more discussion of the banking issues that matter, follow Talking Banking Matters preferred podcast platform.
Roshan Varadarajan, McKinsey: Bhavin, you started Zeta with a focus on a single product, but you’ve now expanded to have a stack that’s one of the leading card-issuing platforms but also covers areas well beyond that. Can you talk a little bit about the founding journey of this? You didn’t grow up specifically in payments or in financial services; you were much broader than that. So what was the trigger? How did you start? And today, how do you create value for your customers?
Bhavin Turakhia, Zeta: Back in 2014, my cofounder and I started thinking about the company, and our vision was what we described as democratizing banking. We believed that a significant amount of financial prosperity could be obtained for individuals and organizations across the world if everybody had access to banking services at fair terms. We looked at the number of people across the globe that still needed access to banking services—meaning the ability to save, grow, and move their money and get access to credit—and we felt we had two possible paths. Either we could become a bank, or we could serve banks. Pretty quickly, we decided that starting a bank would limit us in terms of both return on equity and growth across multiple countries, given the regulatory issues involved. Also, our core competence was more toward software. I’ve spent the better part of my life building enterprise-style software.
Our strategy was to help democratize banking by providing next-gen platforms to banks. We’re not a fintech, we’re not a lending company, and we’re not a bank. We’re a technology service provider to banks, helping them innovate faster. On average, banks take 20 to 30 months to launch a new product.
Our goal was to rewrite the entire banking stack. Not layer on top, not build point solutions, because that’s already the direction we saw most of the industry going in. You either have people hollowing out the core, people layering on top of the core, or people building a point solution. And our perspective was that we wanted to build what we call an omnistack. We strongly believe it has substantially more value, because it basically targets the entire issuing and acquiring stack.
Cycle time is critical to innovation, so it is deeply rooted into our being. The faster we launch, the faster we get feedback, the faster we repeat the cycle. If that cycle takes a year, your compounding growth in ten years is going to be x. If that cycle takes a month, your compounding growth is not going to be 12x, it’s going to be thousands of times higher in comparison, because each iteration compounds incremental marginal improvement. If I can bring that cycle time down to a day, that compounding effect grows astronomically in terms of what value you can derive.
We also wanted to reduce what this costs banks and increase their income across all aspects of banking, including operational, IT integration, and innovation costs. And we increase income by increasing engagement and customer attention. Banks need to delight their customers.
Our goal was to rewrite the entire banking stack. Not layer on top, not build point solutions, because that’s already the direction we saw most of the industry going in.
Bhavin Turakhia, cofounder and CEO, Zeta
Roshan Varadarajan: Can you zoom in a little on what the perfect Zeta customer looks like? What does your target bank look like? And then, for them, how does working with Zeta change the unit cost of, say, launching a card program or running a card program more broadly?
Bhavin Turakhia: Our focus is predominantly on banks, meaning chartered, licensed institutions across the globe that have the right to manufacture financial products and services. We offer a fairly massive technology stack, mostly still on the issuing side. We are still at the early days of addressing the acquiring side. So on the issuing side, we have a fairly massive technology stack. It’s divided into three buckets. The first is our core systems, which cover all of card processing, including credit cards, debit cards, and prepaid accounts. Then there is all of core banking, including demand deposit account savings, term deposits, et cetera. And we have a loan management platform and loan management system that can serve any unsecured loans. We don’t do mortgages yet, but we can do unsecured installment-based loans. Depending on who we target, we offer different products. It might be the whole stack, or it might be a card-in-a-box solution.
In addition to that, we also have Zeta Studios, which is kind of like an accelerator. If you buy our stack, you can also engage a team within Zeta Studios that could be used for a period of time to build out your experiences and integrations, including writing code for your software, as a product development exercise.
Roshan Varadarajan: Let’s dig further into what you described as the core offering. You did a bit of market canvassing in this hot space, where there’s a number of players who, at least on the surface, say they’re tackling the same issues. What are the design choices you have made, both technical and nontechnical, that differentiate you—the structure of the stack, how you think about go to market, how you think about target customers? How do they set Zeta apart?
Bhavin Turakhia: It’s one thing to use next-gen tech; it’s another thing to provide next-gen capabilities. We call ourselves next gen because nobody else provides the capabilities we’ve built. That’s really the way we distinguish ourselves. By doing this, we have the ability to decompose banking entirely, in terms of technology, into its most foundational components—kind of like an operating system, if you will. We call them foundational modules. And then all products are layered on top of this polymorphic omnistack architecture, so Zeta’s fundamental IP [intellectual property] is concentrated in our four foundational modules. Also, we’re cloud native, cloud agnostic, and cloud portable. Underneath our foundational modules is a private cloud layer that allows the rest of the stack to be completely cloud agnostic.
Being headless in this way means we first designed the platform by thinking in terms of APIs, and then UI [user interface]. There is not a single operation of the platform that we need to perform, issuers need to perform, or clients need to perform that cannot be done through an API. The stack is a fully open data stack, which means you have the ability to basically extract all the data in whichever way you want to. We provide reports, APIs, data extracts, data lake, Parquet files, and whatever else you want basically in terms of data.
The most important architectural attribute of our platform is around the potential for personalization, which we achieve through the software principle of inversion of control [where the control flow of a program is managed by a framework or external source, rather than the application code itself]. We built this into every single aspect of the platform, so you can change system behaviors whenever transactions or account-processing steps are taking place. Whether that is authorization, end-of-day processing, or pricing decisions, et cetera, you can essentially specify declarative rules or local JavaScript to control behavior. That gives you unparalleled capability of hyper-personalization of creating creative products.
Roshan Varadarajan: If we look beyond Zeta at the industry as a whole, you alluded to AI and even gen AI levers of value creation, whether it’s on servicing or other journeys. Can you help us separate a little bit of where you’re going to see hype and a lot of action in the near term, versus where experiences or value will be significantly enhanced through gen AI over the next three, four, five years in financial services?
Bhavin Turakhia: I think generative AI is a value generator that is substantially at a higher level than any other technology paradigm that we’ve seen the last two decades, including smartphones, cloud, the internet, et cetera. The compounding effect and the potential for banking are very high. I divide that potential into two buckets. One is decisioning AI, and the second is conversational AI.
For the first, I believe in the next decade, decisioning AI will pretty much run the bank, because banking is by far the most fundamentally conducive industry to be run and operated by AI in a meaningful way. Banking is basically the optimization of a single mathematical equation. The output is return on assets. So that becomes all these decisions: underwriting, pricing, how much interest and fees to charge, fraud decisions. Currently, all these decisions are being made in isolated forms by some kind of basic AI model. Marrying them together and getting to a point where AI can run the bank is the holy grail on the decisioning side.
As for the conversational AI side, I think there is a lot of opportunity to improve the user interface and consumer experience. If you think back, we used to speak to bank tellers to make transactions, and then that changed to apps, but they really haven’t been able to provide as good of a customer experience as humans were able to. Conversational AI has the ability to substantially make a difference for consumer interaction. And if you think about it, conversation is the default for humans; we learn it throughout our life. So it’s the most intuitive way for us to obtain information or perform actions. It’s also not constrained by code. An app is constrained by code. It can only do those ten things that somebody has programmed it to do. With a conversational AI platform that has access to the right kind of large language models, you can generate code on the fly.
Conversational AI has two broad use cases. One is personal agents for the consumer and for personas within the bank—a personal agent for a customer content agent, or a personal agent for an operations person, or a personal agent for a compliance person. The second is customizing offers. This can include marketing the products with the right offers at the right time to customers, to maximize engagement and customer acquisitions, by using transaction history to predict what you’re likely to want to buy and when. And it can include collections. If I provide this hardship plan or connect at this particular time with a particular message, the likelihood of collecting on that debt is higher.
Roshan Varadarajan: You’ve been at this a couple of times now, having started multiple tech companies. They’ve all been in different industries or different subsectors, though. What is the mental framework you use to prioritize problems you want to tackle and figure out where you can actually build a great business?
Bhavin Turakhia: There’s actually a fairly well-defined framework in my mind, and the main reason is that, over the last 25 years, I’ve actually made every mistake known to entrepreneurs that can possibly be made. I had to go through the cycle multiple times, launching products that were insanely stupid, wasting all my capital in investing toward that product, et cetera. I learned a bunch of lessons. So now every new business I operate goes through these four stages: planning, discovery, scaling, and steady state. And each phase has a very clear, well-defined template, structure, deliverable, and outcome as an output which can then serve as a precursor to the next stage.
During the planning stage, the goal is to get conviction around “Does this actually have legs?” You need to understand a clear persona with a clear problem and a proposition that will solve it ten times better than any existing solution. It’s not just a hypothesis you came up with yourself in your head. Now you come to go to market. You could have a great idea, but if you can’t take it to market, then again, the idea has no legs. And there are three pieces to this: pitch, pricing, and promotion. So you have a proposition, but can it be articulated in a manner that is well understood? And what’s the story? What’s the pitch? Does it resonate with the customers you have?
And then can the idea scale? Does it deserve your time in terms of capital, resources, people, and so on and so forth? So you look at it from the top down to try to figure out the total market potential and from the bottom up by projecting the business out some years. Five years is typically what I do for every single business plan, which is a full P&L that shows, “This is going to be my income, my expenses, my net EBITDA,” and what does the business look like in terms of NPV [net present value] in a long-term projection.
Only after all that homework has been done and you get to a point of conviction is when you exit the planning phase and move to the discovery phase. The deliverables in the discovery phase are that you want to put out a minimum viable product that has product–market fit. A second element of the discovery phase is that you want to have at least one traction channel that can scale.
In the scaling stage, the objective is to build playbooks and standardized processes and to increase efficiency on operations and start measuring all the important KPIs. And then after scaling, the eventual goal of every business is to become boring. That’s what I call steady state, which means you get to a point where everything is predictable and growing at a steady clip.
Roshan Varadarajan: Talk to us a bit about your management style. You’re juggling teams across multiple businesses, markets, and products within each of these businesses. How do you think about building, motivating, and retaining phenomenal talent? And what is it like to work at Zeta?
Bhavin Turakhia: I spend a lot of time personally on recruitment and hiring, and we’ve created a culture of figuring out the right fit. It takes time to understand what you’re trying to hire for and figuring out the right individual for that role. But it’s worth spending two times the amount of time to get that right person, because the right person is 20 percent better than what you otherwise might have compromised for. That’s a three- to five-year time frame that will outweigh the extra few days or few months you had to spend in sourcing that individual.
Most of my management takes place through KPIs, especially when I get midway through discovery to midway through scaling. So, each function or organization unit has their own list of critical KPIs. We’re checking progress against KPIs at a regular cadence, sometimes weekly, sometimes monthly. And the goal of the functional head or leader is to keep moving those KPIs in a positive manner. We’re always looking for needle movers and orbit shifters among the business-as-usual types of tasks: What are the needle movers that will meaningfully make an impact on a KPI?
“Our firm is designed to operate as one—a single global partnership united by a strong set of values. We are equally committed to both sides of our mission: attracting and developing a talented and diverse group of colleagues and helping our clients create meaningful and lasting change.
From the C-suite to the front line, we partner with clients to help them innovate more sustainably, achieve lasting gains in performance, and build workforces that will thrive for this generation and the next.”
Please visit the firm link to site