Payment Friction Is Costing Trading Platforms More Than Failed Transactions

Evidence tier: A1 Evidence type: Auto-discovered official publication Source: Rapyd News Official publication date: 2026-08-04 Captured: 2026-08-04T14:08:03.624Z

Rapyd’s David Rosa joins the Fintech Wrap Up podcast to discuss AI Agents, Stablecoins and the Future of Payment Operations

Fintech Wrap Up host Sam Boboev recently spoke with David Rosa, General Manager at Rapyd, about how artificial intelligence, stablecoins and real-time payment infrastructure are changing the way financial businesses operate.

Rosa, who oversees AI Transformation, FX, Payouts and Platform Solutions at Rapyd, discussed the practical use of AI within payment operations, the role stablecoins could play in global dollar movement and why the payment experience has become a competitive issue for online trading platforms. The episode, published on July 10, also covered OpenUSD, autonomous payments, cross-border infrastructure and the effect of MiCA on crypto businesses.

At the center of the conversation was a clear message: payment infrastructure is no longer only an operational concern. It can directly affect how quickly a business grows, whether customers remain on a platform and how effectively the company responds as new technologies change the movement of money.

AI Is Becoming an Operating Model, Not Just a Productivity Tool

Rosa explained that Rapyd is already applying AI to merchant onboarding and internal operations rather than treating it as a future initiative.

According to Rosa, more than 80% of Rapyd merchants are onboarded instantly, while 96% complete onboarding within 48 hours. He said this is achieved while conducting the due diligence required of a regulated financial institution, including validating corporate information and assessing requirements across different jurisdictions and industries.

For sectors such as online trading and brokerage, that process can be particularly complex. Businesses may operate across multiple markets, serve customers internationally and face additional regulatory scrutiny. Rosa said Rapyd combined industry expertise with applied AI to map both the operational process and the judgment required to assess less straightforward cases.

He also drew a distinction between using AI to reduce costs and using it to change how a business operates. In Rapyd’s case, he said the larger opportunity comes from reducing unnecessary layers in decision-making, improving visibility across teams and moving more quickly when new commercial opportunities emerge.

“Cost cutting is really a byproduct,” Rosa said. The greater effect, in his view, comes from accelerating communication and decisions so that the business can grow more efficiently.

Why Payment Friction Costs Trading Platforms More Than a Failed Transaction

The relationship between payment performance and customer retention becomes especially clear in online trading.

Markets move quickly, and traders need access to funds when an opportunity appears. If topping up an account takes days, or in some cases longer because dollars must move through traditional cross-border systems, the customer may no longer be able to act at the right moment.

The cost to a platform therefore extends beyond one failed or delayed deposit. Rosa said customers become frustrated when competing platforms provide a faster and smoother experience, even when the difference is caused by the payment infrastructure operating behind the platform.

“A payment stack is not just an operational efficiency,” Rosa said. “It is an insurance policy against losing clients.”

Better payment performance can also help a platform attract activity from competitors. When users can fund their accounts more reliably and access liquidity more quickly, the payment experience becomes part of the platform’s overall value to the trader.

Local Payment Methods Need Global FX Infrastructure

The way a trader deposits money can differ substantially between Brazil, Poland, South Africa and other markets. Local bank rails and alternative payment methods can make it easier for customers to fund an account using a familiar option, and many of these systems now operate in real time or close to real time.

That does not solve the entire cross-border problem.

Rosa explained that many investments and trading positions are still denominated in US dollars. A customer may begin with Brazilian reais or another local currency, but the platform still needs to convert those funds into dollars and, eventually, convert them back.

Local payment connectivity therefore needs to be supported by a reliable global FX operation. Without it, a platform may improve the first stage of the deposit experience but still encounter friction when funds need to move between currencies or jurisdictions.

This is also why payment strategies cannot simply be copied from one country to another. Local regulations, capital controls, payment habits and currency requirements all affect how money can enter and leave a trading platform.

Stablecoins as a Practical Rail for Global Dollar Movement

Rosa distinguished between stablecoin activity connected to crypto trading and the use of stablecoins within payments.

While much of the market’s existing volume may still be related to digital-asset trading, he said payments are where stablecoins are beginning to demonstrate broader operational value. Their relevance is particularly strong when businesses need to move dollar-denominated value across borders and outside conventional banking hours.

“To me, and by extension to Rapyd, stablecoins equal dollars,” Rosa said.

He was not suggesting that stablecoins will replace the US dollar itself. His point was that dollar-backed stablecoins can provide another way to hold and move dollar value, particularly outside the United States, where traditional dollar transfers can remain slow and operationally complex.

That makes stablecoin adoption a payment-infrastructure decision rather than necessarily a bet on crypto as an asset class. For businesses managing cross-border invoices, payouts or account funding, programmable digital money could provide a more direct connection between 24/7 settlement, FX and automated payment processes.

Rosa identified accounts payable as one area where AI agents and stablecoins could converge. Cross-border invoices are frequently denominated in dollars, while the process of reviewing and paying them remains highly manual. An AI-enabled system using programmable money could eventually evaluate an obligation, apply the appropriate permissions and initiate the payment through a suitable rail.

What Changes When AI Agents Can Initiate Payments?

As payment decisions become more autonomous, the question is not only whether an AI agent can complete a transaction. Businesses also need to determine what each agent is authorized to do.

Rosa connected this challenge to role-based access control within a company. Different employees may have permission to read information, change records or approve specific actions. Agentic payments will require a similar structure, with clear rules governing which agent can access data, initiate a transaction or commit company funds.

The payment stack will also need to account for failures and exceptions. If an autonomous transaction is incorrect, blocked or disputed, the business must know where responsibility sits and how the action can be reviewed.

For that reason, Rosa expects governance to be one of the main factors shaping agentic payments. The underlying technology may allow a transaction to happen without a person in the immediate flow, but businesses will still need control, oversight and accountability around the system.

The conversation also addressed regulation, including MiCA and the licensing requirements affecting crypto platforms in Europe. Rosa’s view was that blockchain-based businesses cannot treat regulation as optional or assume that new technology removes existing obligations. The fiat and digital-asset markets are moving closer together, making a credible licensing and compliance strategy increasingly important.

The Risk of Waiting Until the Payment Stack Breaks

Rosa closed the conversation by warning businesses against assuming that an existing payment setup is adequate simply because it still functions.

“Why fix it if it ain’t broken?” he asked. “It may not be broken, but it is going to be seriously challenged.”

AI agents, stablecoins and new real-time payment systems will not cause established infrastructure to disappear immediately. Rosa expects incumbent networks and payment companies to respond and adapt. The larger risk for businesses is failing to evaluate alternatives until customers, competitors or market conditions force the issue.

That means understanding where payment friction exists today, developing credible alternatives and considering whether a different operating model could be stronger than the current one.

For Rapyd, the conversation reflects a broader approach to global payments: applying AI within operations while connecting local payment methods, FX, payouts and emerging digital rails across markets.

Watch or listen to the full Fintech Wrap Up conversation with David Rosa.

Full Edited Transcript

Speakers

Sam Boboev: Host, Fintech Wrap Up David Rosa: General Manager, Rapyd

Editorial note: This transcript has been lightly edited for clarity and readability. Filler words, repeated phrases and obvious transcription errors, including instances where “Rapyd” was transcribed as “Rapid,” have been corrected. Speaker changes have also been clarified where the original transcript combined the end of a question with the beginning of an answer. The meaning and sequence of the conversation have been preserved.

The first 38 seconds are an edited promotional montage assembled from comments that appear later in the interview. They are labeled separately below. Statistics and market references introduced by the host are reproduced as discussed in the episode and have not been independently verified.

Promotional Preview Montage

[0:00] Sam Boboev: You had to let go of some senior-level executives, highly talented people, because you saw AI doing a really good job. Does that mean AI will take over our jobs in the future?

David Rosa: I think yes, in many ways. AI is going to replace traditional roles.

David Rosa: We wanted to validate whether we could lean on applied AI to replace this.

David Rosa: We are operated by AI. We’re really walking the talk.

David Rosa: To me, and by extension to Rapyd, stablecoins equal dollars.

David Rosa: The card rails are going to be disrupted as well, like everyone else.

David Rosa: Why fix it if it isn’t broken? It may not be broken, but it’s going to be seriously challenged.

OpenUSD and the Stablecoin Market

[0:39] Sam Boboev: David, welcome to the show. You are General Manager at Rapyd. The first thing I want to ask is whether you have been following the latest news about stablecoins and what is happening in the market.

I’m referring to OpenUSD, a new stablecoin launched by more than 140 companies. What was your reaction when you read the news?

David Rosa: First of all, Sam, thanks for having me. It’s a pleasure to be here.

Yes, it is big news. It is very interesting because it is effectively open-sourcing a rail that has proven to add real value.

We have seen a lot of solutions looking for a problem in the blockchain world. Stablecoins have proven that there are real pain points being solved, which I’m sure we’ll get into.

The business models that have been evolving around stablecoins have created depth, but this is very disruptive. It is going out into the open with backing from much of the industry. It seems like a new standard is being implemented and a new chapter is opening in money movement.

[1:52] Sam Boboev: Rapyd also recently launched stablecoin solutions for its merchants and customers. Do you think you will add this stablecoin to your stack, or will you stick to the major ones at the moment, including Circle and Tether?

David Rosa: At the moment, there is no alternative, to be very frank. Although the announcement is huge, there is still some time required to get there.

There is a formal launch, and there is also a regulatory aspect that needs to be addressed globally. That typically does not happen overnight, although the backing is clearly very strong.

Today, it is business as usual. Let’s see how this evolves. I’m sure there will be some very interesting responses from the incumbents, and we are following this with great interest.

Overall, I think the major beneficiary will be the client because there will be another option for moving money in an even more competitive way, 24/7 and around the world.

Stablecoin Volume, Payments and Crypto Trading

[3:05] Sam Boboev: I see massive numbers around stablecoins. Some say that $33 trillion was processed in 2025. When Chainalysis analyzed the market and stripped out bots and internal transfers, it seemed that real-world stablecoin transactions were closer to $9 trillion.

From what you see at Rapyd, working with players in the market and with your merchants, how much of this is actually commerce and how much is crypto trading?

David Rosa: First of all, we are in the world of payments.

From what I have been reading, and anecdotally from my own connections in the market, the brunt of the activity is still related to sitting out market volatility in the crypto-trading world. That is still the main use case.

It solves a pain point, but in terms of making a day-to-day difference in how the world works, it has not yet had the same impact.

Payments are where stablecoins have started to be used and embedded, and that is where we are seeing value. Yes, it is a minority of the overall volume today.

I think having these new protocols and much more competition in the market will push ultimate usage, effectively replacing the US dollar outside the US. That is the main service these stablecoins are providing, in my mind.

[4:47] Sam Boboev: We know that the two major stablecoins in the market, USDC and USDT, are dollar-backed. It seems that perhaps 97% of all stablecoin transactions are tied to the dollar.

Do you really believe that stablecoins will replace the dollar? It does not seem that way to me.

David Rosa: Not replace it, no. To me, and by extension to Rapyd, stablecoins equal dollars.

What I mean is that stablecoins in currencies that already have real-time payment rails, and in currencies that are not necessarily used for global trade, can be solutions looking for a problem.

Take stablecoins in euros when you already have SEPA Instant. Yes, you could enhance the ability to move euros on blockchain instead of through SWIFT, but the brunt of euro invoicing is within the SEPA area and within its own market, so to speak.

The US dollar is a very different beast. Whether we like it or not, it is the truly global currency, and moving it is friction-heavy.

This is not something new, by the way. We can go back to the 1950s, when the so-called Eurodollar came about to allow US dollar deposits outside the United States. It began in London. The pain point has always been there.

That traditional market was challenged because the price of US dollars outside the United States, which was the LIBOR market, was completely disrupted and liquidity disappeared. Stablecoins are picking up where the Eurodollar left off.

That is very interesting. It is not only for payments. It is also relevant to securities in the fiat world. I think it is exciting to see a continuation of that, and it is very much about the US dollar. We are unapologetic about it.

What Rapyd Does and How It Uses AI

[6:54] Sam Boboev: I will dive deeper into that topic in a moment, but first, for my audience, David Rosa is a General Manager at Rapyd. Rapyd is a payments company and one of the leading payments companies in the market.

David is responsible for AI Transformation, FX, Payouts and Platform Solutions.

David, so that my audience understands, what exactly does Rapyd do? How are you different from companies such as Stripe or Adyen?

David Rosa: We are clearly in the business of helping merchants get paid. Helping them connect to their revenue is extremely important and creates a very aligned type of activity.

We do this globally, and we are very present in emerging markets. That is one of the major differentiators in that aspect of the business.

You have probably seen the recent headlines that we have acquired businesses in Latin America, Africa and Eastern Europe. That establishes us even more strongly in emerging markets.

I think the biggest difference is that we are actually doing what many people are still talking about, which is being operated by AI.

If you look at our operations, including the important task of optimizing the customer journey beginning with onboarding, I think we have established a gold standard today.

I challenge people to come and beat us in terms of our ability to onboard merchants. More than 80% of our merchants are onboarded instantly around the world.

That does not mean we are winging it. We are conducting proper due diligence, and we rely heavily on applied AI to do that. Ninety-six percent are onboarded within 48 hours.

It is an experience that is second to none, and we welcome the challenge. In all honesty, it has taken us years to build.

Like everything AI has touched, it has accelerated exponentially during the past 12 months or more. It is highly enabling for our partners and clients.

We are operated by AI. We are really walking the talk. We are also very present in emerging markets. Those are the main differentiators for Rapyd.

Using AI to Improve Margins and Decision-Making

[9:14] Sam Boboev: I think AI is especially useful in emerging markets, particularly in onboarding.

In emerging markets, it is not always easy to collect all the information because the systems may not be as established as they are in the United Kingdom, the United States or other markets. You have to request more details and conduct thorough background checks. I believe AI could be one of the best tools for that.

Rapyd’s CEO has said publicly that AI will triple the company’s profit margins within three years. I want to understand where that comes from in a payments business.

Are you optimizing the business or cutting costs? How are you planning to triple the profit margins?

David Rosa: That is a very interesting question, so let me dig into it.

First, we implemented our “operated by AI” initiative from what we believe is a position of strength because we have reached profitability.

This is not the typical exercise where people say, “Let’s use AI as an excuse to cut costs.” That is not the motivation.

What we have done, and it has been challenging to implement and build support for across the different teams, is create efficiency by having fewer people involved.

I know that sounds crude, but it is human nature. You can have very smart individuals, but once you have one, two or three meetings with many stakeholders, people can disengage very quickly. Ownership of the issue becomes less clear.

As you grow, you can end up with these amorphous meetings. At one point, we grew to more than 1,000 employees, and that creates bureaucracy.

I previously worked in a very large corporation, and it is demoralizing when you see that happen so quickly as a startup grows.

The main point of the exercise was to reduce the number of humans involved in decision-making. Humans are still very much involved, but there are fewer of them, or fewer chefs in the kitchen, as we call it.

Believe it or not, that has been a huge enabler because information flows much faster, as it does in the early days of a startup.

That is empowering for anyone who has been an entrepreneur or is interested in entrepreneurship. All of a sudden, you cut out red tape.

That creates efficiency and helps multiply the top line because you can pursue business much faster. You then combine that with the automation that applied AI provides. I mentioned onboarding, for example, and that turbocharges the process.

Cost cutting is a byproduct. Of course there is some cost reduction. That is part of the reality. But it is not the largest factor in achieving higher margins or tripling those numbers.

The greater opportunity is growing the top line with a multiplier effect because the right kind of communication happens much more efficiently.

Two Applied AI Use Cases at Rapyd

[12:18] Sam Boboev: You lead the AI transformation at Rapyd. Can you walk me through one use case where you implemented AI and it is genuinely working?

I would like my audience to understand a practical use case.

David Rosa: Yes, absolutely. I will give you two use cases.

The first, which we are very proud of, is how quickly we have been able to automate the onboarding of businesses across different jurisdictions.

For any regulated financial institution, onboarding a business can be challenging, even more so than onboarding the individuals behind the business.

You need to access corporate registries and validate the information you have been given. You trust, but you verify.

You also need to augment that due diligence depending on the industry of the particular client.

This is where many competitors or other players in the market avoid certain sectors. Retail trading and online brokerage are examples. Those businesses can be challenging because they are truly global and regulated.

Many providers decide that the sector is too difficult and do not get involved.

We approached it by hiring subject-matter experts from the industry and automating the process. In simple terms, we captured their knowledge and mapped not only the operational process but also the mental process.

We looked at how experts assess edge cases and then automated that process from end to end.

That has allowed us to extend a business model that was operating primarily in Europe across many different geographies.

That is one important example. It did not happen overnight. It required a substantial investment, but it was very worthwhile.

The second example is more complex and extremely interesting because it is at the leading edge of how heavily we are leaning into applied AI.

I mentioned that we have reduced the number of people involved, in some cases at relatively senior levels. These were people who were knowledgeable about particular processes, alignments and refinements.

In product development, for example, you often have synchronous meetings, whether in person or online. Everyone gathers, and very bureaucratic processes can come out of that.

We removed that process completely.

There are functions we have not replaced when people left. In some cases, we let go of some very good people. That had nothing to do with the individuals.

We wanted to validate whether we could lean on applied AI to replace the process.

Instead of using a synchronous process to align and refine the many different moving parts of a company that is constantly building, we created what is effectively an operating system.

The process is now asynchronous, and it gives us what we call forward visibility.

For example, two different departments might be building something with dependencies on one another. The system can flag that, at the current pace, those teams will clash in two weeks because they are not aligned.

That is very empowering. It is also very new.

I think it gets to the heart of using AI to multiply financial and operational metrics.

Will AI Replace Traditional Jobs?

[16:03] Sam Boboev: David, I have to ask this question because you brought up the topic.

You mentioned that you had to let go of some senior-level executives and highly talented people because you saw AI doing a really good job.

Does that mean AI will take over our jobs in the future and affect job security? I have to address this because it is one of the biggest questions in the market.

David Rosa: I think yes, in many ways. AI is going to replace roles that have become traditional today.

But this is nothing new in terms of a paradigm shift.

Depending on which generation you are part of, think about when the internet arrived. I know it may sound like a well-trodden topic, but it is true.

I used to be a trader, and suddenly we had an almost free flow of information. You were bombarded by many different sources. That was highly disruptive, and you had to adapt.

Some parts of the industry effectively sunset. Some old jobs do not exist anymore. But people adapted.

What is brutal is the speed at which this is happening.

The last thing I would say is that, counterintuitively, emerging markets may be affected the most, in my mind.

Before heavy automation through AI, you used to throw people at a problem. If you had an operational problem, it could be outsourced to an emerging-market jurisdiction.

Now the machine can perform that work faster, cheaper and better.

That is a major challenge for service-oriented emerging markets in particular. I am not talking only about countries that manufacture physical goods. Service industries are being disrupted.

Call centers, for example, are already being significantly affected by AI.

AI Agents, Stablecoins and Card Rails

[18:03] Sam Boboev: There is a theory that AI agents will optimize transaction costs and route payments away from card rails, which usually charge 2% to 3%.

The theory is that AI agents will instead use stablecoin rails that can cost a fraction of a cent.

From your perspective at Rapyd, is that a realistic scenario? What are the timelines?

David Rosa: I think it is a realistic scenario, and it disrupts even us, to be very frank.

We are very focused on the agentic-payment topic. It is still a buzzword. Many people are talking about it, but there is relatively little action today. We are looking at it very closely.

My personal opinion is that the part of the payment cycle that is most psychologically painful, and typically most full of friction, is parting with your money.

It is not so much collecting money from your clients or receiving revenue. It is paying your costs.

Psychologically, when you need to pay your costs, you are already cringing because it is not pleasant to part with your money. Operationally, it is also difficult.

I think accounts payable and automation will be at the forefront of this change.

A lot of cross-border invoicing is denominated in dollars, which brings us back to the stablecoin theme we discussed earlier.

You want to do that with programmable money. Why would you want to continue doing what we do manually on a daily basis?

I think that is definitely one area.

The card rails will also be disrupted, like everyone else. The card schemes are waking up to this and are very much at the forefront.

You have seen M&A activity from card schemes buying stablecoin-infrastructure companies, including some very good ones. Those are the initial steps.

I think this will eventually happen, but it will not mean free money transfers around the world.

The idea of “free” can create a race to the bottom initially, and then the market eventually stabilizes.

I would not say that today’s incumbents will simply disappear. I doubt that. But they will be challenged.

We are all going to be challenged, and we are all trying to be on the front foot.

Governance for Autonomous Payments

[20:25] Sam Boboev: Visa and Mastercard are always present when there is a new technology.

For example, Mastercard launched Agent Pay for machines this month. In that flow, AI initiates and completes payments without a human being involved in the process.

What changes in the payment stack when there is no person in the flow and only machines complete the transactions? What would change in Rapyd’s stack?

David Rosa: This is actually a problem we have been dealing with internally.

The topic is RBAC, or role-based access control. Who has the right to access certain information? Who has the right to write information, rather than only read it?

That is an internal pain point because you have many different people across different jurisdictions, with different responsibilities, whom you need to govern.

The lessons we have learned from our own internal governance of data and data privacy are directly applicable to agentic payments.

Which agent will be allowed to do what?

What happens if something goes wrong, which it probably will at some point? This is business, and things go wrong in business from time to time. What happens then?

I think governance across the entire agentic-payments space will be one of the main drivers.

That is how we are adapting and preparing for it, based on our own internal experience.

MiCA and Crypto-Platform Regulation

[22:02] Sam Boboev: If we are talking about governance, MiCA’s transitional period ended on July 1.

Some of the market numbers I have read suggest that approximately 1,200 platforms were operating under temporary registration in Europe, but only around 210 received full authorization.

What happens to the platforms that did not make it, in your opinion?

David Rosa: We have seen several different situations.

The more mature businesses had a clear and robust plan in place. They may still be in a queue, but they have been applying for CASP registration, regulation or licensing in credible jurisdictions.

For them, it can be a question of when rather than if.

When there is a clear direction of travel and the business has prepared itself properly, we tend to work with that. You want to support the business, but the path toward the appropriate licensing needs to be robust and credible.

You are also dealing with government institutions or central banks, which typically move more slowly, especially during the summer months in Europe.

At the opposite extreme, some businesses have been silent and have no real plan or vision. In those cases, you have to tell them to do their homework and come back when they are more prepared.

There are some cases in between, but primarily we have seen those two extremes.

It is very disruptive because there are now serious consequences. You cannot improvise a regulatory setup.

It does not matter whether you are operating in the blockchain world or the fiat world. It is still regulation.

In fact, the two worlds are coming together. These are interesting times. July 1 was showtime.

How MiCA Affects Pay-Ins and Payouts for Trading Platforms

[24:00] Sam Boboev: You work with trading platforms at Rapyd. This is your vertical, and MiCA creates a new compliance layer for anything touching crypto.

Do you think this will affect pay-ins and payouts for crypto platforms moving forward?

David Rosa: It already has.

Even after moving past the July 1 deadline, I think we will continue to see the primary friction that has affected this industry.

First, it is still a very misunderstood industry.

I used to be a trader, so I find it easier to understand and connect with the language. Once you are aligned with the terminology, you can start making progress.

More importantly, think about the experience as an investor. You want to get the timing right.

Timing is important in life generally, but it is especially important when you are trading in financial markets.

You can see an opportunity coming, but then you realize that you need to top up your account. It can take days, and sometimes weeks, because US dollars are moving through SWIFT.

That is the pain point.

The regulatory setup is very important, as we discussed, and I believe that will be worked through over time.

But the real pain point is the need for instant gratification in an industry that genuinely requires it.

Markets move in the blink of an eye. You want liquidity to be available so that you can express your view on the market.

That is what we specialize in.

The True Cost of Failed Deposits

[25:40] Sam Boboev: Most trading platforms accept that some percentage of deposits will fail. They budget for it, write it off and move on.

In reality, what does that cost a platform when you consider the complete picture, rather than only the failed transaction?

David Rosa: It adds up substantially.

You are absolutely right. I think it costs clients.

People become frustrated because some platforms do this better than others.

When you look under the hood, you realize that it is often the underlying payment stack that enables a better, smoother and more immediate experience.

A payment stack is not just an operational efficiency. It is an insurance policy against losing clients.

In fact, it can also help you grow because you can capture activity from competitors when you provide a much smoother way for clients to express their views on the market.

Local Payment Methods and Conversion

[26:40] Sam Boboev: Trading differs from country to country. A trader in Brazil deposits differently from a trader in Poland or South Africa.

How much does offering local payment methods in each market change the experience and improve conversion rates?

David Rosa: It does make a difference to some extent, assuming that the platform and the type of activity are properly licensed and covered from a regulatory perspective in those emerging-market jurisdictions.

Do not forget that many emerging markets have capital controls in place.

As a result, much of the activity has a local dimension, although in some cases you do have cross-border transactions.

When you move into cross-border payments, the reality is that you are back in the world of US dollars.

Connecting local payment rails helps. Increasingly, payment systems in emerging markets operate in real time or close to real time, whether through banking rails or alternative payment methods.

But you then need a highly efficient FX operation.

The investment or market position that the customer wants to take is often denominated in US dollars.

Even if the customer begins with Brazilian reais, those funds need to be converted into US dollars and then converted back at the end.

Local payment methods matter, but if you do not complement them with a strong, truly global FX platform, you have a solution looking for a problem.

Building Rapyd’s FX Infrastructure

[28:08] Sam Boboev: What does the FX stack look like at Rapyd? Can you give me more information?

David Rosa: This was one of the first things I built when I joined Rapyd.

As I mentioned, my background is in trading, and I was deeply involved in emerging markets, which effectively means a lot of FX.

I used to joke with Rapyd’s founder and CEO that we did not know how to spell FX at Rapyd.

The mistakes I have seen, even in banks and among professional traders, include inverting bids and offers or writing lines of code in the wrong direction.

It is extremely easy to make very basic mistakes.

We reduced the operation to strong but fundamental building blocks and educated the entire company about how to “spell FX,” literally.

I went on an internal campaign to explain and build understanding around FX.

It paid off because people were interested. They knew FX was part of operating a global business.

That was about four years ago. Since then, we have continued to rely on that strong foundation.

You either have that foundation or you do not.

Many businesses eventually experience a serious incident because of an incorrect decimal or because bids and offers have been inverted.

This is something we invested in very early, and it is part of the stack we offer.

Why Real-Time Domestic Payments Do Not Automatically Work Cross-Border

[29:44] Sam Boboev: You briefly mentioned local real-time payment rails.

Brazil has Pix, the United Kingdom has Faster Payments and India has UPI. We have fast, real-time payments domestically.

Why has that not translated into cross-border payments?

When I am trading, I want to pay out very quickly from one country to another, but it does not work that way. I still have to wait. It is not the same as using Faster Payments in the United Kingdom.

David Rosa: In my mind, it is because there is no central government authority that acts as a global clearing house.

These payment systems move money, and money is at the heart of a jurisdiction, its legal system and its hierarchy. It is ultimately connected to what creates the legitimacy of a nation or jurisdiction.

Once you begin crossing jurisdictions, you also encounter geopolitics. Some countries do not like dealing with one another.

How do you create a neutral, truly global platform that operates 24 hours a day, 365 days a year and helps clear those payments?

You do not have one.

As a result, you have private-sector platforms. Rapyd is one of them.

These platforms try to aggregate as much volume as possible, ideally with two-way flows, so they can net the activity and capture the bid-offer spread in the middle.

The reason there is no central clearing system connecting all of these real-time rails ultimately exists at the government level.

I also think this is where stablecoins work very well because you can combine them with on-chain FX.

When we discussed the pain points stablecoins solve, the traditional model was sometimes called a sandwich structure. You on-ramp in one jurisdiction and off-ramp in another, with everything happening in close to real time.

Now people refer to atomic settlement. It is a new term for a similar outcome.

Stablecoins lend themselves well to that structure.

With a new protocol such as OpenUSD entering the market, it could turbocharge these types of platforms. I am very excited to see how it develops.

OpenUSD, Libra and the Importance of Governance

[32:17] Sam Boboev: OpenUSD reminds me of Libra from Facebook.

In 2019, Libra launched with many major names, including PayPal and others, but it ultimately did not work out.

Perhaps the outcome will be different now because regulation is in place.

As you said, there are still many different opinions in the market about governance, so I am curious to see how this evolves.

David, this is my final question. I want my audience, especially merchants, to take away one key point from the discussion.

You work across FX, payouts, platforms and AI at Rapyd. What mistake do you see businesses making repeatedly when it comes to payments, and what would you recommend they change?

The Biggest Mistake Businesses Make in Payments

[33:13] David Rosa: The mistake people make, especially now that applied AI is beginning to affect industries, is asking, “Why fix it if it ain’t broken?”

It may not be broken, but it is going to be seriously challenged by all of these major changes.

The mistake is like driving a car while looking only in the rearview mirror.

You need to invest the time and effort required to stay abreast of developments.

These developments can challenge us as well. Our clients are also looking at their Plan B and Plan C.

You need to prepare a Plan B and a Plan C. Those alternatives could eventually be substantially better than Plan A.

The mistake is mental inertia.

You see a lot of that in finance generally, which is an industry full of intermediaries.

This is also why the applied-AI approach I mentioned earlier, removing unnecessary layers and involving fewer people, allows the signal to move much faster.

I encourage people not to become complacent.

Do the research, cut through the red tape and stay abreast of developments because the world of payments is changing very quickly.

[34:27] Sam Boboev: David, I really enjoyed the conversation. Thank you for coming on the show.

David Rosa: My pleasure.