When looking to the future of something like B2B payments, it is best to focus on just one aspect and from that, you start to see a numerous other areas of change coming into focus. You don’t have to look to far ahead to identify one very clear trend and
driver in the payment’s world, and that is an ever increasing drive towards real-time payments that can be made at any time, 24x7x365.
First off, let’s understand that expectations of business banking are heavily influenced by what we as consumers can do/services we receive. Equally, consumer focussed expectations flow directly into business expectations, and that's simply because businesses
are made up of consumers, made up of people.
As consumers, our expectations are driven by technology companies, not banks. It's companies like Facebook, Amazon, Snapchat and WhatsApp that set our expectation level, and that's because we now access so much of our life online via connected devices, such
as laptops, tablets and mobile phones. If my experience across these applications is real time, with real time feedback loops and status alerts, then that's what I expect.
For example, with WhatsApp I send a message and I can see if it's waiting to be sent, or that it has been sent by my device. I can see when it's delivered, in real-time. I can even see the recipient come online and read the message, with the status 'ticks'
turning blue. That's all in real time. So, when I pick up a mobile banking application, and I want to send money, I expect naturally the same sort of experience as I get from apps that I use constantly throughout the day.
When Monzo launched its initial offering, you wouldn't say it was as mature as product offerings provided by its incumbent bank competitors. However, they tapped into the simple fact that as users, we have an expectation of real time and real time feedback.
The experience of paying on my Monzo card, to then get a real time alert down to my phone, was pretty much unique.
Then opening the app to see the transaction was already debited from my account was almost unheard of. What Monzo did great was to simply tap into that real-time experience and expectation of their customer base. Businesses are no different, they too have
an ever-increasing expectation that payments can be made in real-time, any time and enjoy instant feedback loops.
Businesses and real-time
Businesses can send money in real-time using FPS in the UK for example, however the amounts can be limiting, and even when FPS can be used, that feedback loop just isn’t there. Business banking is not as simple as consumer and business needs are quite different
even between businesses, but that desire to have less friction in the payment process is consistent.
Businesses do not want to be chasing payments, either because they haven't arrived or because reconciliation is a struggle. Likewise businesses want to make payments as late as possible to improve cashflow, and with larger sums, massive sums even, they wish
to hold on to that money for as long as possible, earning that free money in interest.
Businesses and their finance teams also want that experience of being on the phone to a supplier and be able to pay them there and then, with the supplier being able to confirm they’ve received it there and then too. Therefore, real time payments with real
time feedback loops are inevitable. This need for real-time though is not a geographic specific one, rather it is seen globally in all jurisdictions. Cross-border payments therefore should be seen no different.
Cross-border should be no different
The cross-border payment experience should be no different to a real time domestic payment. As the globe continues to wake up to global economics, businesses find that their supply chains are increasingly becoming international, and not just tied to a single
geography. Cross-border possess harder challenges than domestic to satisfy real-time, namely that of currency liquidity, visibility and the fact there isn’t a payment system operating 24/7.
Payment friction is paid by someone in the process, and that someone is often the businesses involved. Payment friction removes transparency visibility into cashflow and what’s paid and unpaid. That for many businesses means they delay making a payment themselves,
delay investment, delay kicking off a new project, delay servicing another customer. Some will borrow to smooth out that cashflow, costing the business, others may even purchase exotic products to try and smooth out their FX rates due to “pre-funding” cross
border payments days if not weeks and months in advance. At a global scale this friction costs trillions of dollars and locks up business liquidity unnecessarily. The impact on productivity is unknown.
You don't need to use technology for technology's sake to solve these problems. Blockchain is often touted as the way to solve these issues, but actually they introduce risk associated with the value of the transaction. Right now, millions of payments reach
their destination with a value that is not what was expected by the recipient, or by the party initiating the payment.
This is because of the unknown route, the admin fees associated and how many stops there are on their way to the final destination. This is all problematic, and we’ve not even mentioned FX rates, or the fact that these rates will change between a transaction
being initiated and its fulfilment. There are also costs baked into the system, including trying to provide transparency of where a cross-border payment is in the chain.
Businesses will need to have real-time, just-in-time cross border payment capabilities, and many expect this type of service.
Technology stepping in
Technology is moving at an ever-increasing pace, if anything moving further away from industries like banking and insurance faster than ever before. Technologies like cloud solutions - following event pattern models and platforms sitting on top of global
infrastructure - are enabling engineers to deliver new levels of capabilities. These capabilities and outcomes mean businesses are becoming less patient with financial systems in solving these problems.
If we look at the cloud, we have two massive global providers making up most of the “cloud” infrastructure, with a third entering the market. Technology is now easier than ever to deploy into specific geographies, and PaaS is enabling small agile technology
focussed companies to build powerful solutions quickly. We have seen a number of payment service providers start to leverage these capabilities, and the cross border challenge has seen a wealth of new FinTech companies bring slick solutions to market, players
like TransferWise, Transfermate and EMQ, all with similar models, all with similar challenges but all focussed on solving one main problem. Real-time delivery of a transaction, and with that visibility and feedback.
Innovation in this area of real-time looks set to increase, with the majority of central banks ensuring real-time payment systems are being built and implemented. And, in the cross-border space we have new FinTech entrants every few months it seems, including
the venture I am part of, RTGS.global.
Real-time then will happen, and that’s very much the future for B2B payments. But what do we learn from this? Well, enabling real-time isn’t easy. This means that operational systems within banks, payment service providers and even line of business applications
will have to change. In some cases moving away from batches will cause real technical challenges and migration nightmares for some.
This means systems have to embrace event based patterns, understanding horizontal scale and distributed data across their systems. For many, this will prove too far a leap to make, and for those that fail, they will lose market share quickly. However, once
you’ve solved this enablement part, providers must ensure payments process all the way through automatically, straight through process (STP) raises its head.
STP may sound simple, but actually payments go through various and numerous steps on their way to get processed, more so in the world of cross-border. The challenge is set then to make sure as many transactions as possible process automatically without human
intervention. This sounds simple, but it really isn’t. Even if you have 99% STP that 1% could cost you millions when we look at real world business payment volumes.
For modern systems the reason for a transaction not to STP is almost always related to either some nuance in the payment payload, or, related to financial crime screening.
The FinCrime challenge
In our drive for real-time, we've now seen that the future of B2B payments includes investment in internal systems, controls and now FinCrime screening, all to ensure STP rates support real-time expectations and experiences.
When we were building ClearBank, we were always aware that STP was key to our banking capabilities and BaaS product offering. We invested heavily in a fraud detection system and built our very own sanctions screening capabilities, all because we wanted to
drive those STP rates up and up, ensuring our customers could deliver real time experiences to all their customers.
The challenge with FinCrime comes down to basics, trying to identify real financial crime activity or participants in financial crime, and getting better at removing false positives. False positives is what kills STP rates for payment providers. Predicting
a 'match' is a false positive and processing it quickly and effectively therefore becomes a focus point. Technology such as deep machine learning has a real role to play at this point.
First off, can machine learning be used to understand the profile of a business, the type of payments it makes, the values, how they are initiated and in context to other businesses making payment? It most certainly can. It also can be used to reduce false
positives with techniques such as matching transactional and activity profiles across businesses. By learning here and reducing the number of false positives, focus can be placed on actually identifying FinCrime actors. This type of learning is predominant
looking at B2B payment fraud.
Secondly, when we look at transactions themselves, can machine learning be used to learn which payments or actors are false positives and which are of genuine concern, specifically around sanctioned businesses, vessels, individuals or countries? The answer
is yes. This again means transactions that are identified can be worked by an operator.
Thirdly, when looking at how an operator identified transactions for potential FinCrime, could we again use machine learning to speed up this process. The answer is yes. Machine learning with a feedback loop from operators can build out powerful capabilities,
with the machine more accurately predicting that the identified transaction is actually a false positive. With greater data sets our machine learning gets better, and that means we have less friction in the system and only additional learnings that further
help the machine. It also means that as a payment provider we are spending time on real FinCrime and not impacting B2B payments.
Trusted digital business identity
The real kicker here though in delivering greater STP, and ultimately that real time experience for all, is re-establishing trust in the payment parties. If, as a payment, the payment service provider and bank they trust the entities, then there is no screening
of the transaction. This used to be very much the case for domestic to domestic FPS transactions, but in recent years, that trust model has been proven to be broken. So a trust model doesn’t work if you can’t trust it, therefore B2B payments' future will end
up with a trust model that’s is actually built on 'knowing'. 'Knowing' is only possible if we have true digital identities being used as the payee and the beneficiary.
With a true digital identity that can be used to initiate transactions, transactions can move to an identity paying an identity. These both exposing claims about themselves, including such things as holding structures and UBOs. This can be shared with the
sending and receiving banks and therefore decision making on 'if to execute' the transaction becomes a digital binary decision.
In doing so, our AI and machine learning technology is now focussed on fraud screening and learning behaviours of identities. This is far more focussed, far more accurate and with a true digital identity we move our STP rates ever closer to 100%. At worst
here, we have instant feedback loops and status updates, so that customer experience is real-time no matter the outcome.
This is all technically available today. The challenge is not the technology, it's the culture to drive change through rapidly.
The impact of IoT on real-time
Staying with our singular prediction of real-time payments we start to explore IoT. If we believe payments will be real-time, have ever decreasing levels of friction, then we must also believe that payments will get initiated by businesses automatically,
and therefore by connected devices within the business.
An example makes this clear. A production line, manufacturing cars, consists of numerous robots and robotic type arms. These are becoming more and more IoT devices in their own right. These devices understand when the robot needs maintenance work, or if
there is a malfunction of a particular part. These devices can now not only diagnose their own issues but initiate the ordering of new parts or schedule in their own repairs and in doing so, trigger a payment. These devices are essentially ensuring just in
time smaller payments that need only be reconciled as opposed to initiated at all by the finance team.
As more and more smart devices become connected the more activities that they initiate will involve a payment. IoT drastically alters the payment landscape, raising the volume of real time transactions dramatically, which in turn puts far greater pressure
on payments providers for high straight through processing rates. In addition, at an infrastructure level, there are now a new set of requirements evolving, predominately those associated again with digital identity, authentication and payment reconciliation.
There is much made of the relationship between payments and identity. However, in the case of IoT this becomes crucial, as too does the relationships between identities, something that often gets overlooked. With IoT devices, we need to identify not only
the device, but also the company that the device is registered too, its location (if it is ordering parts etc) and the identity of those individuals who can 'sign off' on any work/the actual payment. We soon have an identity relationship tree that provides,
ultimately, down to the device a form of delegated authority to carry out these tasks and initiate these specific payments to specific beneficiaries.
Identity and authentication for a payment service provider therefore become heavily interlinked. The big issues here are clearly those linked to security and trust, however, in terms of the payment itself, it’s the ability to carry some of this identity-based
This is not a case of simply adopting the ISO 20022 LEI format, rather the ability to have a true identity with a set of trusted claims attached to it, this time an IoT device. This level of information is needed by businesses initiating the transaction,
and it's great for the payment service provider to help with authentication security and clearly for audit capabilities. The identity information may also help with reconciliation for both the business initiating the payment and receiving the payment.
With IoT based transactions, the issue for a finance department is now the additional volume of smaller transactions and the additional challenge of reconciling those payments. It’s clear that Payment Service Providers will have to modify how they expose
transactional based data and for smaller businesses, provide different tools for not only exporting transactions but querying transactions.
Greater data will also be required for transactions to be reconciled by those businesses which are the beneficiary of a specific payment. The payment infrastructure will have to evolve to meet these growing and changing demands of business to ensure the
efficient processing of payments by the payee and beneficiary.
By focussing on one macro-level driver of change within the payments world, in this case the increasing drive towards real-time payments, we can see a cascade of change and evolution that impacts not only businesses, but the people working within those businesses,
payment service providers/banks and their operational processes, underlying technology and technical capabilities. We can even start to predict the necessary changes and requirements for the underlying payment system infrastructure.
There is without doubt an ever-increasing demand for real-time payment capabilities. These capabilities need to be available all of the time, and this requirement itself will drive great changes within banks and payment service providers. Greater investment
in financial crime tools, machine learning and operational efficiencies to ensure high STP rates will all be areas of change.
IoT will further distort the current payment profile of a business and with that, how we identity actors, authenticate payments, enable delegation of authentication and reconciliation. Digital identities will therefore become increasingly important, having
a role to play in FinCrime, payment STP capabilities and reconciliations.
These changes in the future of payments aren’t specific to a single geography or region, rather these will be global changes and trends, therefore cross-border payments must be treated in exactly the same way.
There is a lot of change coming to the B2B payments space, all of which is highly exciting and beneficial to businesses and their customers. Banks and payment service providers need to be well positioned to deal with the challenge of real-time, and the ripple
effect it will have on their business going forward.