New Core Alternative: Runs in the Cloud, Built by a Bank
As head of research and development at the largest retail bank in Ukraine, Alexander Vityaz grappled with a problem faced by banks around the globe: old, inefficient IT systems that hindered data processing and product innovation.
The solution he came up with — a core platform run entirely in the cloud — worked well enough for PrivatBank to send him to Silicon Valley to try to sell the technology to the U.S. banking industry.
Corezoid, the vendor spun off from PrivatBank, officially opened for business last week with a sales pitch that might have been a nonstarter a few years ago but sounds plausible today as banks race to keep up with customers' expectations.
PNC recently held an API Fest where it asked its employees to pitch customer innovation ideas. With the move, PNC joins a growing number of U.S. banks embracing the possibility that open APIs offer.April 29
"If there are 30,000 banks in the world, they're all trying to create 30,000 different Internet banks, 30,000 antifraud systems, and so on," Vityaz said. "Ninety-nine percent of business operations in the bank are the same standard. These operations need to be commoditized and moved to the cloud."
U.S. banks' reticence to use cloud services for core functions will soon be a thing of the past, some observers predicted.
"Typically there have been the twin concerns of security and protecting customer data and regulatory compliance" discouraging banks from using the cloud, said Jim O'Neill, an analyst at Celent. "I think security concerns will soon vanish; you saw the same thing in the early days of mobile banking, and now every bank offers it. And I think the regulatory fears are based on presumption. I don't see anything in the FFIEC guidelines that calls out cloud computing as not acceptable for banking services."
Also, the current practice of adding application layers on top of those old systems will soon prove impractical, said Julien Courbe, a financial services and technology analyst at PwC.
"Adding digital integration layers is a bit of an awkward workaround," he said. "The evolution in cloud-based solutions will make [cloud] a more appealing solution to banks."
Corezoid, though a bit different in what it does, joins other vendors aiming to sell banks on putting core processes in the cloud. These include Nymbus, a cloud-based core processing vendor based in Miami Beach, Fla., and Wilmington, N.C.-based nCino (which was also originally part of a bank). Also, longtime financial technology vendor Misys announced last month that it would move its full product suite — including core systems — to the cloud.
Like many institutions, PrivatBank, based in Dnipropetrovsk, Ukraine's third-largest city, required enormous computing capabilities to power the millions of financial transactions conducted per day at the bank. This hurt operating efficiency (when the bank needed a quarterly report, a server couldn't quickly process the huge amounts of information necessary) as well as customer experience (systems couldn't process data quickly enough to offer a customer a real-time view of their account balance).
To solve these issues, Vityaz — with a $3 million budget — created what is now Corezoid in 2013. He started by compiling a list of all the kinds of banking account calculations that might ever be necessary, then developing algorithms to handle each of these.
Essentially, he said, this cuts out the hardcoding needed to create banking services, instead mixing and matching application programming interfaces, or APIs, to customized algorithmic processes.
If-this-then-that formulas are matched to customers' needs and situations at a given moment. The result is a new "digital core" enabling banks to use a single cloud platform for all of their processes, said Vityaz, who spoke through a translator.
Further, Vityaz said Corezoid acts as a kind of "Krazy Glue" for banking APIs. "Right now, each bank has an IT department inefficiently reinventing the same bank-specific solutions," Vityaz said. "With Corezoid, banks can instead use ready-made processes … simply by connecting the APIs they need. That's the promise that cloud banking has to offer."
So, account and deposit processing are now performed at PrivatBank through Corezoid's cloud operating system. Once a customer makes any change in their account — like adding or removing funds — that operation is sent through Corezoid's algorithm-based process engine, which finds in its system a predefined process associated with this account and type of transaction, and completes the necessary calculations instantly. If Corezoid doesn't find the necessary account template in the system, it automatically creates a new process, Vityaz said.
In addition to PrivatBank, Corezoid recently picked up its second client: Western Union is using Corezoid as the back end of its new online money transfer service in Ukraine. Corezoid is now a separate company from PrivatBank, though the latter is an investor in it. Including Vityaz, the company has 20 employees, all of whom previously worked at the bank.
As Corezoid decided to look for a permanent headquarters and begin offering services on a larger scale, it chose Silicon Valley (Redwood City, Calif., to be precise) in part to be able to reach the U.S. banking industry.
"We needed to talk directly to this market," Vityaz said. "The problem is that American banks are trying to be IT companies. They would be much happier if they could concentrate on the main banking business instead of being IT companies."
Corezoid's offering is perhaps not so much a traditional core banking platform per se, but something more akin to a Swiss army knife allowing banks to potentially solve specific problems, O'Neill said.
"It seems like a general-purpose toolkit for building workflows — a platform for building custom apps," he said. "A bank might have some data and processing issues and effectively use this kind of service to build an app to solve these issues."
Core banking itself will be a driver of cloud adoption among banks, O'Neill said.
"For banks, their typical tools are mainframe-based," he said. "Now, they are very good at what they do, to scale to millions of transactions, and that's what they have been using for over the last 50 years. But they're not very agile in terms of creating new products and services and having [the ability] to adapt to new business models on the fly."