Manhattan-based FinTech Current’s use of graph database technology is key to its approach to building new financial services for customers: creating a set of ‘hybrid finance’ products based on integrated views of individuals and their family connections.
Brought in before the brand’s 2017 launch, graph’s ability to handle interconnected data is praised by the firm’s internal IT team, with claims that the data modeling and performance advantages have helped it win millions of account holders since.
Those wins started in conventional financial services like checking and savings accounts, but a new integrated model of a current customer is also now helping it move into non-traditional, blockchain based ‘DeFi’ (decentralized finance).
Since opening for business, Current, which styles itself a ‘financial technology platform,’ not a bank, was quick to offer potential customers a range of products.
These include a debit card, basic and premier current accounts, cash back, money management tools and support for Apple Pay. But it’s the way it can connect these products with customers and non-customers is the bank’s USP, said the brand’s Chief Technology Officer, Trevor Marshall.
This is where graph - sourced from supplier Neo4j - makes its contribution. He said:
When we were orienting our data model and the way we built around our concepts, we knew this was our future. A graph made total sense, as it allows us to expand our ability to acquire customers and roadmap in terms of the products our customers want. Neo4j is critical to making that work, as our data model can change if we add other products and the architecture we’ve defined is very flexible. We often define new nodes to add to the user graph. Neo4j allows us to model anything.
Graph came to his attention when he and his team were looking at ways to support multiple record lookups and allow complex database queries. That was needed, as Current’s proposed model was to be based around connecting members to the best experiences to enable them to improve financial outcomes.
Achieving that meant building a completely new core customer database, which unlike traditional banks, would be centered around a user, instead of their separate accounts.
This would be the best way to innovate and build products faster, said Marshall, plus offer the ability to connect to both traditional and decentralized backends.
Building a new kind of account structure
Relational databases didn’t seem the right match for doing that, as running a query to find all the relationships and qualified signatories for a minor account would take too long due to the constant need to create new tables in-memory. as JOINs are executed in SQL. This would cause performance issues.
Another need was tight integration with the FinTech’s chosen cloud platform, Google, where Kubernetes is being used, as well as native Google technology like Google Pub/Sub and Dataflow.
Marshall added that he chose this supplier because of the ability to create new types of models through relationships, as well as ease of checking the access a user can add or delete new products (i.e., modify certain parts of the graph version of their customer record), which is important for creating an easy to use, fast and reliable experience.
Finally, Marshall liked the base graph data paradigm, as he believes it maps the data model to the way that the data is actually architected in the database. That makes it very clear for any developers who join the team to quickly understand the significance of a given node, the relationships that node makes visible in a customer record, and how that all comes together.
Graph also prides itself on surfacing useful relationships in data. This was in fact the basis of the first Current product, a checking account banking for teens. To get one, an adult family member needs to sign up for a Current account, then link to a new one for younger family members. To ensure a seamless experience and enable a parent to add multiple teenagers (or anyone else) to their account, Marshall says graph was the best option. He explained:
Graph helped us build a new kind of account structure through a householding graph. Using Neo4j, it is possible for a parent, or grandparent, or anyone to add more people to their Current account in a flexible way more aligned with modern family structures.
This form of software also allows explicit mapping, allowing the Current core banking engine to traverse to different data elements very quickly.
As a result, he said, developers can enter the Current graph through many points and find relevant data with simple queries. Graph structures also give Current flexibility in how it integrates with banking partners, he said.
Since the product’s debut in 2017, Current now has nearly four million members, which Marshall sees as “the culmination of what we’ve been building over the past seven years.”
Now, graph is now being used to aid the company move to DeFi—the alternative payment processes for trading in cryptocurrency or obtaining a mortgage without recourse to an intermediary using blockchain-based smart contracts. The aim at Current now, he said, us to bridge traditional and decentralized finance to build what Current frames as ‘hybrid’ financial products, which it believes provide the most value back to our members in the shift to Web 3.0.
Finding connections in data
Marshall said the logic for the move is fully aligned with the organization’s original ethos of bringing premium financial services to everyone:
We have already proven we can serve our customers in a trustworthy way through our traditional banking products. Now, we’re bringing them value in a way that prevents the extreme technical understanding required to extract that value directly.
Hybrid finance is all about connecting and bridging, which is what we have wanted to do since the beginning at Current.
That move has already begun, with a partnership announced last year with a decentralized finance platform built on the Polkadot blockchain network called Acala, which will eventually lead to it introducing in-app DeFi tools.
And again, using graph as the basis for a holistic customer record is useful, as the tech’s ability to find connections in data is expected to be highly relevant for its new DeFi work.