If you’ve bought something online through an online retailer or through a merchant’s web site or app and you’ve selected to buy now and pay later with interest free instalments, a micro-loan is created for you. That capability of a business to offer you financial products at the Point of Sale (PoS) within their own experience without the consumer needing to get in touch with a Bank is basically Embedded Finance. Buy-Now Pay-later (BNPL) is an embedded lending use case and is probably the most common one in retail.

Embedded Finance is not limited to retail of course. Other interesting use cases can be found across Business & Corporate Banking, Private Wealth, Insurance, etc. For example e-commerce platforms can offer merchant cash accounts, payment processing, etc to SMEs i.e merchants. The list goes on. Embedded Finance is basically unlocking the financial landscape for banks to sell their products at the point of sale within the experience of a non-bank. That’s a massive addressable market for Banks with some estimates of up to $7 trillion by 2030.

The evolution of the Point of Sale (PoS)

With some of the most interesting use cases in retail, the definition of the PoS starts to blur when embedded finance kicks in.

The virtual PoS is typically found at the checkout phase on a retailer or e-commerce website or app where you can select your payment method. This is where most of the time consumers are offered with financial products such a BNPL loan, a new credit card, etc. On the other hand, to most of us a physical PoS is a device at a retailer where we tap a card, a phone or a watch to complete a payment.

Building a virtual PoS within a website or an app to offer financial products is pretty straightforward. You just have to write the code. For example, if a retailer wants to offer BNPL at their website, all they have to do is add a button of a BNPL provider such Paypal or Klarna at their checkout. That is, in oversimplified terms, a module in their web or mobile app that will integrate with the BNPL providers APIs that most likely are deployed somewhere in the Cloud.

Every connected device can be a PoS device

But what about a physical PoS? These are still around. Most of the time traditional physical PoSs are cashiers machines or apps linked to a card reader. They have no ability to “offer” financial products to a customer at the checkout as frictionlessly and seamlessly as the respective virtual PoS experience. PoSs have started to take many shapes and forms however, especially with the explosion of the Internet of Things (IoT). For example, Apple has started rolling out the Tap to Pay capability which turns an iPhone into a PoS.

What would it take for a supermarket to offer a BNPL plan during the checkout to their loyal customers for example? If you think the idea is crazy, Tesla can offer insurance premiums based on the drivers real-time driving behaviour. Voice assistants can make orders. Checkouts are everywhere. Every connected device can be a PoS device.

Imagine having to originate a new financial product at a traditional retailer PoS. That can’t happen when your PoS is a button-first device or its hardware can only run software that’s limited to scanning RFID chips with NFC or generating QR codes. The PoS device needs to adapt as needed to meet the constantly changing requirements of the financial product they originate. To make that happen, we need to treat the PoS device more like an IoT device who’s software can be updated over the air.

Bringing Banking closer to the customer with 5G, IoT and Edge Computing

IoT is not new. As an emerging technology, like many others, it is evolving rapidly by two main drivers. 5G and Cloud Computing. 5G is providing a network with faster and more reliable connectivity, while cloud platforms bring the computing services required to capture the large amounts of data IoT devices generate, analyse and decide in real time.

IoT coupled with 5G and Cloud computing can turn a car, a voice assistant or a smart supermarket self-service checkout machine into a potential embedded finance PoS. That sounds very interesting and one can only imagine the potential. There’s only one challenge though. And that’s with the cloud.

Public cloud platforms have greatly evolved over the last 10-15 years. The big hyper-scalers like AWS and Azure are continuously developing their wide range of data centers across the globe to include more and more geographies and countries bringing cloud services even closer to the source of data. The caveat to this model is centralization. Cloud data centers still are centralized in specific locations and not distributed. That proximity gap is filled in by another emerging technology, edge computing.

Edge computing’s idea is to bring computing even closer to the source of data. Its a paradigm that relies more on distributed architecture than a monolithic one. Although most cloud providers at this point provide some sort of distributed computing services like distributed DBaaS, etc, these do not scale out to the edge network. AWS and Azure are on the path to develop their edge capabilities.

Distributed Banking Architecture over the Edge

With the use of edge computing services the need for data to be transmitted over long distances can be greatly reduced. Latency or the delay between the time when a request is made and the time when a response is received is minimised. For example, if a consumer is using a smart supermarket PoS or a car to make a financial transaction, edge computing can be used to process the transaction at a nearby edge server rather than sending the data all the way to a central server for processing. This can make the transaction faster and more responsive, which can be particularly important for time-sensitive financial transactions such originating a loan at checkout, or making an instant payment.

The ripple effects of embedded finance, edge and cloud computing, IoT etc are multi-dimensional. One thing is certain. These are layers of technology that at the very core require Banking Capabilities (Lending, Deposits, etc) that can be deployed at scale in a distributed architecture.