In this FinTech Interview, Ran Grushkowsky, CEO of MassPay shares his thoughts on how AI-powered orchestration is reshaping cross-border payments into instant, compliant, and frictionless global payouts.
Ran, could you share your professional journey and the experiences that led you to co-found MassPay?
I’ve spent my entire career watching distribution outrun money.
In the late 90s, I was experimenting with digital distribution in the MP3 era – well before Napster took off. We watched content move across borders in seconds while monetization lagged on clunky, domestic rails. Later, building remittance and payout products at companies like USend and ATMCash, I was staring at the ugly reality of cross-border flows: correspondent banks, broken data standards, opaque fees, and last-mile partners that could derail an entire transaction because a single field was formatted incorrectly.
The pattern was always the same: users expected instant, intuitive experiences, but money still behaved like it was 1985. People had accepted that “this is just how cross-border works” and stopped questioning it.
By the time I co-founded MassPay, I had built enough infrastructure and lived through enough regulatory and operational pain to know two things: first, cross-border payouts were structurally harder than almost any other problem in fintech; and second, if you could make them feel as simple as sending a message – without cutting corners on compliance – you would unlock an enormous amount of trapped value for platforms and for workers globally.
MassPay is the product of twenty years of pattern recognition and a clear conviction that payouts deserved their own purpose-built orchestration layer.
You’ve built multiple fintech and digital platforms over the last two decades. What patterns or recurring inefficiencies in global payments pushed you toward solving cross-border payout orchestration specifically?
The biggest inefficiency I kept seeing was that global payments were architected around rails, not around outcomes.
Everywhere I looked, systems were essentially hard-wired: money went from A to B over a preferred rail, but if that rail failed or underperformed, the whole flow broke down. That mentality made sense when you were operating in a single country on one or two dominant networks. It falls apart when you’re paying thousands or millions of recipients across hundreds of corridors, each with different rails, regulations, and preferences.
I also saw three recurring pain points. First, brittle routing – one-size-fits-all paths that couldn’t adapt in real time when a rail went down, a bank changed its requirements, or a recipient needed a different payout method. Second, data chaos: every country, bank, and scheme wanted different fields and formats, and a single typo could send a payment into limbo for days. Third, compliance bolted on at the end – screening and KYC/KYB treated as a gate rather than a native part of the flow, which created friction for good users and still didn’t reliably keep bad actors out.
Payout orchestration is the answer to all three. It asks: for this transaction, corridor, and recipient, what is the optimal route across all available rails, given cost, speed, risk, and user preference? And it makes that decision dynamically, in real time, with compliance embedded – not bolted on. Once you see the system that way, it’s very hard to go back to static routing.
MassPay describes its mission as enabling instant, compliant, global payouts. In practical terms, what was the biggest technical or regulatory bottleneck you had to overcome to make “instant anywhere” actually work?
The real bottleneck sits at the intersection of technology and regulation: local reality.
“Instant” is easy to promise at the API layer. It’s incredibly hard to deliver when you’re dealing with local rails that go offline without notice, banks that batch instead of clearing in real time, jurisdictions where KYC, tax, and FX rules change mid-quarter, and recipients who might prefer a card today, a wallet tomorrow, and a bank account next month.
The core breakthrough for us was to stop treating compliance and local constraints as an afterthought and model them as first-class inputs to routing. Our orchestration engine doesn’t just ask “which rail is fastest?” – it asks “which rails are even legal and appropriate for this corridor, use case, and recipient profile right now, and among those, which path gives us the best combination of speed, cost, and success probability?”
That meant building three things in parallel: deep, direct connectivity into local rails and last-mile partners; a rich, constantly updated policy layer that encodes regulatory rules by market; and real-time decisioning that can switch rails in milliseconds when something fails or becomes non-compliant. Getting that triad right is what turns “instant” from a marketing claim into an actual operating characteristic.
The idea of “eliminating economic geography” is powerful. How do you define economic geography in today’s digital economy, and where do you see its most damaging effects on workers and businesses?
Economic geography, to me, is the invisible tax you pay for where you happen to be born or operate.
In a world where work, creativity, and distribution are increasingly global, your zip code still dictates how fast you get paid, in what currency, with what volatility, how much of your earnings leak out in fees, FX spreads, and delays – and whether you can even access the tools global platforms take for granted.
The most damaging effects show up in two places.
For workers – gig workers, creators, freelancers in emerging markets – it means doing the same work as a peer in the U.S. or Western Europe but waiting days or weeks to see the money, sometimes losing 10-20% to intermediaries in the process. That directly impacts their ability to smooth cash flow, invest, and build financial stability.
For businesses, especially platforms with a global footprint, economic geography turns payouts into a drag on growth. It complicates expansion into new markets, increases support costs, and creates churn when earners can’t trust that “you’ll get paid” actually means you’ll get paid now, in a way that works for you.
Eliminating economic geography doesn’t mean ignoring local law or context – it means abstracting the complexity so that a worker in Lagos, a designer in Manila, and a developer in São Paulo experience payouts with the same reliability and immediacy as someone in London.
MassPay uses AI-driven orchestration to route payments across multiple rails and providers. How does this orchestration layer work in practice, and what makes it fundamentally different from traditional payment routing systems?
Think of orchestration as a real-time logistics engine – but for money and compliance.
In traditional routing, you might have a static priority list: try Rail A, then Rail B, then Rail C. That logic is often hard-coded, updated infrequently, and blind to what’s actually happening on the ground.
Our orchestration layer looks very different. For every transaction, it evaluates a comprehensive set of factors in real time – spanning performance, cost, recipient preferences, risk signals, and regulatory constraints. Based on that, it scores potential paths and selects the one with the highest expected success and utility. If something degrades mid-flow – a rail goes down, a partner starts timing out, a compliance rule tightens – the system can re-score and re-route on the fly.
AI and machine learning help us continuously refine those decisions. Every success and every failure feeds back into the model. Over time, the system learns which combinations of conditions predict friction in a given market and proactively avoids those routes.
Fundamentally, we’re not just moving funds – we’re orchestrating a portfolio of rails and partners as if they were one adaptive network. That’s a different category than traditional routing. It’s closer to an operating system for global payouts.
You’ve mentioned reducing payment failure rates significantly compared to legacy systems. What role does real-time decisioning play in achieving this level of reliability?
Real-time decisioning is the difference between “we’ll find out tomorrow if it worked” and “we can adapt before anything breaks.”
In legacy setups, a payout often follows a predetermined path. If a field is slightly off, a local rail is under maintenance, or a bank tightens controls, the payment fails – and you only discover that hours or days later. Then you enter the world of manual investigations, support tickets, and re-tries.
With real-time decisioning, three things change.
First, pre-flight checks: before we ever hit a rail, we normalize and validate data against corridor-specific patterns and historical error codes. A lot of potential failures are prevented at this stage.
Second, dynamic path selection: because we have live telemetry on success rates, latencies, and partner behavior by route, we can prefer paths that are currently healthy and avoid ones that are degraded – even if they looked fine on paper yesterday.
Third, intelligent failover: if a payout does encounter friction, the system can quickly classify the failure – data issue, compliance block, rail outage – and either auto-correct or re-route to an alternative rail or partner with a high probability of success.
The result is fewer dead-ends, fewer stuck-in-transit payouts, and a much tighter feedback loop. For large platforms, shaving even one or two percentage points off failure rates translates into thousands of support tickets avoided and, more importantly, earners who don’t have to live with uncertainty about when they’ll see their money.
How are partnerships with networks like Visa, Ripple, and others shaping the evolution of MassPay’s infrastructure and what does an ideal global payouts ecosystem look like in five years?
We approach partnerships with a simple lens: every new rail or network should expand what we can do for our clients and their recipients – without forcing them to re-architect their own stack.
Networks like Visa, major banks, real-time payment systems, wallets, and newer crypto and stablecoin rails each bring different strengths – coverage, speed, cost efficiency, 24/7 settlement, or resilience in markets where traditional infrastructure is thin. Our orchestration layer is designed to sit above all of them, abstracting away those differences so a client can say “pay this user, optimally” and let us determine whether that means a push-to-card via Visa, a domestic real-time credit, a local wallet, or a compliant on- and off-ramp via a digital asset rail in a specific corridor.
In five years, I think the most successful ecosystems will look less like a few dominant networks and more like a dense mesh of specialized rails stitched together by orchestration layers. Clients will integrate once into an orchestration platform and gain access to an ever-evolving portfolio of rails – without being locked into any single one. Compliance and risk rules will be encoded and shared at the network level far more seamlessly than they are today.
Our role in that world is to be the connective tissue – continuously integrating new rails, optimizing across them, and giving platforms a single, intelligent interface to all of it.
Gig workers, creators, and freelancers in emerging markets are a major focus. What specific friction points do they face today when accessing earnings, and how does MassPay’s model change their day-to-day financial experience?
Earnings access in emerging markets comes with very concrete pain.
Access is often delayed – workers can wait days or weeks after a platform marks a payout as complete. Income is fragmented across multiple wallets, bank accounts, and informal channels, so assembling usable funds becomes its own job. Costs and FX rates are unpredictable, with many earners discovering only at the end how much of their pay has been eroded by fees and spreads. And in some markets there is outright exclusion – people who don’t fit neatly into legacy risk models are simply left out of the system.
MassPay changes that experience in three ways.
First, speed and predictability: payouts are orchestrated over the fastest viable rail for that corridor and recipient profile – often in near-real time – and platforms can communicate accurate expectations because we have real data to back it up.
Second, local relevance: we support the payout methods that actually work for those users – traditional local bank transfers, wallets or cash pick-up where banks are less accessible, stablecoins where local currencies are volatile – all through the same unified layer.
Third, transparency and reliability: because we’re monitoring and optimizing success rates continuously, earners see fewer failures and fewer “where is my money?” situations. Platforms can show fees and timelines up front, which builds trust.
When you’re living payout to payout, those differences are not cosmetic. They determine whether you can pay rent on time, take on more work, or stick with a platform long term.
As embedded finance becomes more widespread, where do you see the line between “payments infrastructure” and “financial intelligence platforms” converging in the next decade?
The line is already blurring – and I think it effectively disappears over the next decade.
Historically, payments infrastructure answered one question: “Can we move money from A to B?” Financial intelligence answered another: “Should we move this money, in this way, at this time, under this risk and compliance posture?” Those lived in different systems, often with humans stitching them together.
In an embedded world, that separation becomes a liability. Platforms need payout flows that are simultaneously context-aware – understanding who the user is and how they behave; policy-aware – incorporating what regulations, tax rules, and risk models permit in that moment; and performance-aware – choosing which rail is actually optimal right now.
That’s not just plumbing. That’s intelligence.
Our view is that payments infrastructure and financial intelligence converge into a single orchestration layer that continuously reasons about cost, speed, compliance, and user experience for every transaction. Data from payouts feeds back into underwriting, fraud, and product decisions – and those decisions, in turn, directly shape how payouts are routed.
Finally, for founders and fintech leaders building in cross-border payments or embedded finance today, what is the most important principle or strategic mindset you would recommend they adopt early on?
Respect the complexity – and then lean into it as your moat.
Cross-border and embedded payouts are not just another API. They sit at the convergence of messy local rails, evolving regulation, heterogeneous user expectations, and real-time risk. If you underestimate that, you’ll build something that works in a demo and breaks at scale.
The founders I see winning in this space are ruthless about a few things.
They choose depth over breadth – focusing on the hardest, most underserved parts of the stack and committing to build real infrastructure rather than a thin abstraction. They treat compliance as infrastructure – baking KYC, KYB, AML, tax, and capital controls into core system design rather than treating them as after-the-fact checkboxes, which slows you down early and dramatically accelerates you later. And they prioritize long-term relationships over short-term hacks – investing in direct local partnerships, data feedback loops, and orchestration logic that can absorb new rails and rules over time, even when that work is expensive and unglamorous.
A quote or advice from the author: “At the end of every transaction is a person. All of the infrastructure, the orchestration, the rails and routing logic – it exists for one reason: to make sure that person gets paid on time, in full, with dignity. That’s the only metric that ultimately matters.”
