Fintech: Breaking Down the Barriers

Fintech has become more than just a buzzword. Major banks and multinational corporations have top executives focused on it, looking at ways the growing throng of solutions out there may be able to help their businesses. AFP has taken note of the growing fintech revolution, launching AFP Mindshift at AFP 2017—a host of sessions and events that explore how new technology relates to treasury and finance.

In this latest Treasury in Practice guide, underwritten by Kyriba, we will take a deep dive into these emerging technologies and how they relate specifically to corporate treasury practitioners. The purpose of this guide is to filter out all the noise, providing you with insights into how these innovations can and will impact your profession in the years ahead.

Fintech, for me, just means making stuff work that frankly should have been working five to 10 years ago.”

Change is coming

As Magnus Carlsson, AFP’s manager of treasury and payments pointed out in a recent blog post, fintech development should be aimed at making real issues better for the end-user. Change doesn’t come easily for treasury professionals, so if they are going to adopt any fintech solution, it needs to solve a specific problem.

Indeed, earlier this year, during a meeting of the AFP Annual Conference Planning Task Force, multiple practitioners discussed how, to them, “fintech” means practical solutions. “Fintech, for me, just means making stuff work that frankly should have been working five to 10 years ago,” said a treasurer for a major hotel chain. “It’s getting a treasury management system that talks to the banks, or getting SWIFT connectivity to work. I don’t know why we’re still suffering through that, and I imagine others are too. But day-to-day, that’s what the fintech challenge is for us.”

Another hospitality treasurer agreed, explaining that the treasury workstation his company invested in years ago is now outdated and he is looking at cloud-based solutions. “So fintech for us is whether we can move to the cloud,” he said. “That’s what we’re in the middle of right now; looking for applications that will make things easier and more efficient. So it’s kind of the basics of treasury.”

But fintech is forcing banks to think differently about how they do things. And eventually, corporate treasurers will gradually follow suit. As Carlsson explained in a separate blog, “corporate interest in this space will only continue to grow, since the promises of the new technology can result in great efficiencies at the corporate level. Typically, corporate practitioners on the whole tend to take their time when it comes to changes—especially in terms of adopting new technologies. But we are going to get to the point where the mainstream corporate won’t be able to ignore this growing industry.”

Blockchain and DLT

Although blockchain/distributed ledger technology (DLT) is not synonymous with fintech, you’d be hard pressed to have a fintech discussion without someone bringing it up. And for good reason.

Perhaps more than any other piece in the fintech puzzle, blockchain/DLT has a throng of evangelists ready to tell you that it will completely reshape the way payments and financial services work in the near future. And they may be right—but it’s important to remember that these are still the early days.

According to upcoming AFP 2017 speaker Don Tapscott,
we are entering into a “fourth industrial revolution”—a new paradigm in technology with a new modus operandi for banking and payments. At the heart of this new paradigm is blockchain. “Blockchain represents nothing less than the second era of the internet,” he said. “It can be at the heart of a new financial system if we do it right and we time it right.”

Payments

Anything of value can be stored, transacted, settled, recorded and managed on a blockchain or distributed ledger. For payments, blockchain provides a way to cut out the middleman in cross- border transactions. There are no three-day settlement periods with these types of payments because the payment and the settlement are the same activity. “I think this is the way to get to real-time,” Tapscott said.

But while some blockchain gurus believe it has the potential to be used as the central infrastructure for a national or international payments system, doing so would be a huge, costly undertaking. That’s why the majority of efforts currently underway to establish real-time capabilities in domestic or international payments systems involve traditional payment rails and legacy systems. Additionally, there are questions about whether it could even handle the scale of something like the U.S. economy.

The most mature blockchain currently in operation is the bitcoin blockchain. There are just over 16 million bitcoins in circulation, but the blockchain protocol caps the ceiling at no more than 21 million bitcoins in play at a given time. Now compare that to the more than 600 million traditional payment transactions occurring every day in the United States.

If a distributed ledger were to be used in this type of capacity, it would have trillions of entries in a couple years. “If you are using a distributed ledger for recordkeeping and reporting purposes, will it be manageable for searching, indexing and identifying data? That’s a big, open question that I haven’t seen anyone address yet,” said Daniel Wood, attorney with Pillsbury Winthrop Shaw Pittman LLP.

But the potential for blockchain may actually be in providing information around payments rather than the payments themselves. “Perhaps they might allow for a registry that banks can connect to get KYC information and other information that may have some relation to a transaction but isn’t the transaction itself,” said one real-time payments expert at the recent Payments Innovation Alliance meeting in Berlin.

It’s important to note, however, that for something like that to work, a permissioned ledger that only provides full access to a select few parties would be necessary. Corporate and banks are unlikely to get on board with a permissionless, bitcoin-model ledger that provides open access to transaction data. “Even if we’re entering an era of open data, you still need to have rules and have that trust,” said the payments expert.

Blockchain represents nothing less than the second era of the internet. It can be at the heart of a new financial system if we do it right and we time it right.”

Security

Security is one area where blockchain/DLT has the potential to be revolutionary. Virtually all companies today use traditional server technology for their data, and all of it can be hacked. As we just saw in the WannaCry ransomware incident, failing to download a patch in time can make your entire system susceptible to dangerous malware.

But blockchain provides a way to secure transactions that is much more difficult to hack than what we use today. Each “block” that makes up the “chain” references the previous block. So in order to hack into a particular transaction, “I’d have to hack that block, plus the entire history of blocks on that chain—not just on one computer using the highest level of cryptography but across millions of computers, all simultaneously, while the most powerful computing resource in the world is watching me,” Tapscott said. “I’m not saying it’s unhackable but it’s infinitely more secure than the kinds of computing systems we have in our banks today.”

Nevertheless, it is possible for a hacker to break into a blockchain. “For proof-of-work blockchains, which are generally public blockchains like Ethereum or bitcoin, it is possible to attack those networks using a very large amount of processing power,” said Dan Robinson, product architect for Chain. “That’s one reason why a lot of permissioned blockchains, like what we use at Chain, generally don’t use proof-of-work. Instead, we use signatures by a federation of trusted or partially trusted signers, to ensure that there’s only one version of a consensus at a particular time. This is the approach that a lot of permissioned blockchain companies and platforms are using.”

Trade finance

Probably the most well-documented case of a company using blockchain to solve a problem is the Microsoft/Bank of America solution to improve trade finance transacting.

In 2015, Marley Gray, principle architect program manager for Microsoft Azure Blockchain Engineering, began working with Microsoft’s treasury department, familiarizing them with the technology. From there, they began brainstorming over pain points that blockchain might address. “We settled on standby letter of credit (SLOC),” he said. “We wanted to pick a use case that has business value and isn’t just blockchain for blockchain’s sake, and also one that proved what we assumed was the true value of the blockchain network.”

Gray’s team took a deep dive into treasury’s standard process for getting a SLOC when one is needed. It’s a lengthy ordeal that takes about five business days, on average. “There’s roughly about 14 to 15 steps involved in that process, and of those steps, there’s about a 50 percent error rate with your data,” Gray said. “That leads to a lot of manual reconciliation—phone calls, faxes, emails, etc. It’s just a mess to try and reconcile that so it can be measured between one to two people per counterparty, with a cost estimate per letter of credit of anywhere from $1,500 to $15,000.”

This process is often just between four counterparties— Microsoft, its bank (BofA), the customer and the customer’s bank. But more counterparties can be involved, resulting in an even bigger headache for treasury.

“So we built the system, and we wanted to track its key performance indicators—how long it takes, how many steps, what it costs, human error rate, etc.,” Gray said. “We got it down to minutes instead of days. We got it down to four steps, we got both the error rate and the human involvement down to zero, and we got the cost down to pennies on the dollar. It’s very seamless the way it operates.”

Microsoft then took things a step further and added another counterparty to see what would happen. “That top line—the steps, the error rate, etc.—all of those numbers go up when you add counterparties,” Gray said. “But the blockchain cost line stays flat. So that tells us that the value of the network increases exponentially by the number of participants in that network. Everyone benefits from the lowering of that cost, freeing up capital and speeding up business. So that’s what we proved in the proof-of-concept—it does have business value.”

Bank fee analysis

Another area where blockchain may have potential for treasury is in the bank fee analysis process. During the latest meeting of AFP’s Treasury Advisory Group, Tom Hunt, CTP, director of treasury services for AFP, floated the idea after reading up on Microsoft’s trade finance solution.

“I saw a parallel with banking; it’s very decentralized, and you have a lot of different players and different relationships,” he said. “Why couldn’t we explore the possibility of using blockchain for account analysis?”

In the U.S. alone, there are more than 2,500 AFP service codes. There are so many codes because there is no consistency in the ways banks charge for services. The onus is typically on corporate practitioners to do all the analysis, and it’s typically through manual processes. For large corporates with a multitude of banking partners, that’s a nightmare.

That’s why Hunt believes a blockchain solution could help. AFP is currently in the process of putting together an exploratory meeting with banks, practitioners and blockchain experts to discuss the pain points and see if all these different groups could find some common ground.

However, there would be some hurdles such a solution would have to overcome. Hunt noted that bank fee analysis typically gets “the last tier of technology support” at banks. As such, banks may be reluctant to invest in new technology to improve the process.

Furthermore, revising the service codes and bank fees would be a huge undertaking on the bank’s part. That’s not likely to happen unless there is a major demand from corporate clients. That will require a lot of work on the corporate treasury professionals’ part to pinpoint areas where the analysis process could be improved. So if practitioners want to relieve themselves of this headache—via blockchain or otherwise—they need to put in the time to do it.

Barriers to adoption

But for blockchain/DLT to truly catch on, there needs to be interoperability between competitors, noted Christopher Mager, CTP, managing director and head of global innovation for BNY Mellon. Currently, there are a number of digital ledgers that are emerging, such as Hyperledger’s Fabric, R3’s Corda, Ripple, and more. “You’ve got a lot of ledgers out there that don’t talk
to each other,” he said. “Until interoperability occurs, or one of them emerges as the leading code base, that’s going to impair the network effect.”

The other missing piece is regulation. Thus far, regulations around blockchain and DLT have only applied to digital currencies. Regulators have yet to tackle distributed ledger as a book of record. This creates a legal quandary: Is settlement that happens on a ledger final and legally binding? “That hurdle has to be overcome before you’ll see large scale enterprise applications with a distributed ledger underpinning them,” Mager said.

Regulators in the U.S. and the UK have largely taken a wait-and-see approach to blockchain and DLT, given that they don’t want to stifle innovation. However, some governments have been a bit more proactive. “Dubai and Singapore are the two countries where they’ve embraced the technology throughout the whole ecosystem—banking, corporates and the government,” said James Wallis, vice president, payments industry and blockchain, global industries for IBM. “So in Dubai for example, one of the proofs of concept was around trade. So it was two banks, an airline, a shipping company and the port authority trying to figure out what a new ecosystem might look like.”

Once standardization, regulation and interoperability are sorted out, the use cases for blockchain/DLT are potentially endless, Mager also sees some exciting prospects, such as a convergence of technologies. “The Internet of Things and DLT have a lot of potential overlaps,” he said. “Smart contracts and artificial intelligence have a lot of potential overlaps. I think you’re going to see a convergence of these technologies emerge in the coming years.”

APIs

The use of application programming interfaces (APIs) in treasury is a fairly recent development, largely following the adoption
of the Revised Payments Services Directive (PSD2) in Europe. The directive’s goal is improving competition in banking and payments in the European Union. But since the implementation, APIs have begun to find their way into corporations and banking outside of Europe.

“PSD2 mandates that banks open up their platforms through APIs to payment services providers and account information service providers to be able to initiate payments or provide account information,” said Bob Stark, vice president of strategy for Kyriba. “That’s where it all started; it forced banks to publish APIs and connect to technology providers that want to specialize in initiation and delivery of payments. Many expect this will open the door to new ways of sending payments.”

Real-time processing

Some banks see open APIs as differentiators in that they provide faster connectivity mechanisms than what is currently in use. APIs have the potential to move corporates beyond the batch process for payments, eventually allowing for a real-time connection.

Moving away from batch processing and into real-time processing would be music to the ears of many treasury professionals. “A common complaint for treasury teams has been that they want real-time visibility of banking activity within their cash and treasury management platforms,” Stark said. “APIs offer the opportunity to make intraday reporting into real-time reporting.”

The thirst for this information varies among organizations.
But oftentimes banks will only offer first and second presentment in their intraday reporting. For treasurers to check on particular transactions, they may have to actually go into the bank portal because the bank isn’t providing that visibility quick enough via BAI or MT reporting. But API connectivity will offer more, once banks are in a position to leverage its potential. “That’s where it becomes more interesting for treasury—not just for bank reporting, but also for instant visibility into payment confirmations and eBAM processes,” Stark said.

SWIFT gpi

SWIFT’s global payments innovation (gpi), which went live earlier this year, relies on APIs to provide banks’ corporate customers with faster, more transparent and traceable cross-border payments. A special tracking feature provides treasurers with a real-time view of their transactions, which includes confirmations when payments are credited to recipients’ accounts. Additionally, gpi’s transparency ensures that remittance information is transferred unaltered to recipients.

“The tracking service is a cloud-based service which uses ISO 20022 APIs to reach all the banks in the chain,” said Harry Newman, global head of banking of SWIFT, during a session on gpi at a recent conference. “It enables you to track, end-to-end, a transaction.”

At the same session, Todd Roberts, senior vice president of enterprise innovation for CIBC stressed the importance corporate treasurers place on visibility. “Corporates expect greater certainty and information about the transaction,” he said. “Greater information about the process is unquestionably important to them in terms of how they manage their business.”

Of course, like anything, adoption of gpi will take some time. Roberts believes the biggest barrier to corporate adoption is the complexity of integration. To that end, SWIFT and its bank network need to have an eye on connectivity between parties, using efficient, modern infrastructures like APIs to facilitate that connectivity.

Machine learning and AI

The interest in machine learning and artificial intelligence (AI) continues to grow, but overall, treasury has notably lagged behind in this regard. According to Stark, this is simply because they don’t know enough about it yet. “When treasury better understands the options offered by machine learning, they will become more interested,” he said.

And what can machine learning do for treasury? One example is fraud detection. Machine learning is ideal for pattern recognition, which is essential when trying to spot suspicious activities. “Machine learning is ideal for building complex and evolving rules to detect suspicious activity before it happens,” Stark said. “The sophistication of fraud schemes is increasing at such a pace that machine learning tends to be the only solution to keep up and minimize risk.”

You’ll be adapting your forecast based on a huge number of variables—more than any human can digest. It could be everything from forward-looking environmental variables to projected payment patterns of your customers in different regions.”

Liquidity and cash forecasting

Once corporate treasury professionals become more familiar with business intelligence (BI), they’ll become more comfortable with using machine learning to improve analytics and decision-making around liquidity and cash forecasting.

Let’s say you build your cash forecast in a very static way. You’re looking at what you know is going to happen based on what your controllers in different regions and your treasury staff are telling you. Then you look at the historical information you have, and you build an analysis on that. From there, machine learning comes into play. AI can learn from past data and build that into the analysis, making better predictions about what’s going to happen in the future.

“The computing that goes into that is tremendous,” Stark said. “You’ll be adapting your forecast based on a huge number of variables—more than any human can digest. It could be everything from forward-looking environmental variables to projected payment patterns of your customers in different regions. It could be supplier analytics—what your supply chain is supposed to look like, based on a number of micro and macro variables. It’s taking all of these things into account and then building a forecast that learns from them.”

The key to treasury adoption will be providing practitioners with these practical examples. “When treasurers and CFOs hear about real-life, practical business uses, their interest level leaps forward,” Stark said.

Treasury evolution

However, some trepidation on treasury’s part when it comes to AI could be due to the potential threat it poses to the department in the long run. As we see more and more, traditional treasury responsibilities are becoming increasingly automated. If the bulk of treasury’s work ultimately becomes handled by computers, then many treasury positions could ultimately get phased out, warned Johan Nystedt, vice president, treasury for IR Conagra Brands, during a panel session at the recent AFP Executive Forum.

However, Stark is a bit more optimistic. He believes that some of those treasury roles will simply evolve rather than become obsolete. “There are roles in treasury that won’t exist in 10 to 15 years,” he said. “We’ve already seen some examples of outsourcing in treasury, which was unheard of just a few years back. But look at the ways in which the roles within treasury have already adapted in the past decade alone.

“Cash Managers don’t spend their time logging into banks and manually downloading files anymore. This concept of manual entry has begun to evaporate with the TMS, because it provides automation. It’s that same type of evolution that we’ll continue to see. Things like building your cash and exposure forecasts are going to evolve away. And you’re going to have treasury people analyzing what this all means and how they can utilize that assessment to do something with capital markets, borrowing, etc.”

Furthermore, automating processes also frees treasury up to become more strategic, which is where the function is already headed. The 2017 AFP Strategic Role of Treasury Survey found that treasury is playing a more central role in corporate decision-making, which is where Nystedt believe treasury needs to be. “In the future, it’s going to be less of us being treasury professionals and more of us being partners to the business,” he said.

And as treasurers become part of that decision-making process, they may ultimately find themselves as the ones calling all the shots in terms of incorporating new technology. If so, it only benefits practitioners to keep up with the latest trends in fintech. Otherwise, they run the risk of being
left behind.

As Tapscott said, “Leaders of old paradigms have problems embracing the new.” Treasury professionals may be set in their ways, but they cannot afford to be dinosaurs. Technology is moving too fast, and treasury with it. The time to embrace the future is now.

Key Takeaways

• Change doesn’t come easily for treasury professionals. If they are going to adopt any fintech solution,
  it needs to solve a specific problem.
• Blockchain and distributed ledger technology (DLT) have the potential to cut out the middleman in cross- border payments, while making the transactions more secure.
• Blockchain/DLT has been successful at improving the arduous standby letter of credit (SLOC) process for treasury, and may be able to streamline bank fee analysis.
• Application programming interfaces (APIs) have the potential to move corporates beyond the batch process for payments, eventually allowing for a real-time connection.
• SWIFT’s global payments innovation (gpi) has a tracking feature that provides treasurers with a real-time view of their transactions, including confirmations when payments are credited to recipients’ accounts.
• Machine learning/artificial intelligence (AI) can vastly improve analytics and decision-making around liquidity and cash forecasting.
• AI and automation may pose risks to the viability of certain roles within treasury—meaning that the roles themselves will have to evolve.