Nicolas Pinto

Nicolas PintoAug 27, 2026

AI in Banking Is Moving From a Feature to a New Operating Model 💡 There is something slightly misleading about the way we currently talk about AI in banking. We tend to look for the visible applications: the chatbot that answers questions, the model that detects fraud, the algorithm that improves credit scoring, or the assistant that helps an employee find information faster. These are useful applications, but they can make the transformation look much smaller than it really is. The more profound change is not that banks are becoming better at using AI. It is that AI is starting to change how a bank can actually operate 👨‍💻 For decades, banking has been built around products, processes and organisational silos. A customer has a need, identifies the appropriate financial product, enters a predefined journey, provides information, goes through a series of controls and eventually reaches a decision. AI creates the possibility of reversing that logic 🤖 Instead of asking customers to understand the structure of the bank, the bank can increasingly understand the customer's objective and orchestrate the relevant products, decisions and processes around it. A consumer does not necessarily think: “I need a refinancing product.” They think: “My monthly financial position is becoming difficult to manage.” This is where agentic AI could become particularly significant for financial services. The value will not simply come from generating an answer or making a recommendation. AI will increasingly be able to interpret intent, gather context, assess possible actions, interact with multiple systems and, within clearly defined boundaries, execute parts of the journey 🤝 The competitive advantage may no longer sit primarily in the individual product or even in the interface. Models themselves are also likely to become increasingly commoditised. The harder thing to replicate will be everything surrounding the model: the quality and depth of proprietary data, the context available to the system, the decision frameworks, the workflows connecting different parts of the organisation, and the feedback loops created by millions of real-world interactions. In other words, the real asset may become the intelligence layer sitting across the bank 🧠 A bank cannot simply give an AI agent a goal and allow it to operate without boundaries. Decisions involving credit, payments, fraud, compliance or customer money require a very different approach to autonomy. It will be about designing a system in which AI can reason and act while remaining constrained by policy, regulation and accountability. AI reasons. Policy constrains. Systems execute. Humans govern. This is the difference between adding AI to banking and building banking around AI. Source: Preeti Basu - https://lnkd.in/ezd8PyGe #Innovation #Fintech #Banking #FinancialServices #AI #GenAI #Agents #Agentic #Strategy #Governance

Nicolas PintoAug 26, 2026

Stablecoin Ecosystem 💡 Most people think stablecoins are just “digital dollars.” They’re not. They’re becoming an entire financial ecosystem 👨‍💻 Look closely at this diagram and you’ll see something important: the stablecoin economy isn’t built around a single issuer or a single blockchain. It’s a multi-layer financial infrastructure. 🔹 Issuers Crypto-native issuers operate across multiple blockchains. At the same time, banks are entering the market with their own stablecoins, backed by reserves and traditional banking infrastructure. 🔹 Banks They provide the rails between traditional finance and stablecoins: on/off-ramps, reserve custody and connections with foreign banking partners. 🔹 CeFi Exchanges, liquidity providers and OTC desks create the market infrastructure that allows stablecoins to move, trade and find liquidity. 🔹 DeFi Pools provide liquidity, lending and yield, increasingly becoming another layer of financial infrastructure rather than simply a crypto-native experiment. 🔹 Custodians & wallets Custodial infrastructure, unhosted wallets and cross-border custody connect users and institutions to the ecosystem. 🔹 Payment services And this is where things get particularly interesting. B2B ↔ B2B. Merchant ↔ Retail. Primary markets ↔ Secondary markets. Stablecoins are increasingly moving closer to the actual movement of money. This is why I don’t think the most interesting stablecoin story is about trading. It’s about settlement 💰 Over the next 3–5 years, we could see: → Bank-issued stablecoins competing with crypto-native issuers → Foreign custodians connecting directly to payment infrastructure → DeFi pools evolving into institutional liquidity hubs → Stablecoins embedded directly into B2B settlement → Real-time FX becoming part of stablecoin transfers → Cross-border trade settling without traditional correspondent banking And that leads to the biggest shift of all: Stablecoins moving from trading tools → settlement rails. Once stablecoins become deeply integrated into B2B payments, the conversation changes. It’s no longer simply about “crypto.” It becomes about financial plumbing 🌐 Banks won’t necessarily disappear. They may simply start moving money in a fundamentally different way. Source: TRM Labs - https://lnkd.in/erjYUGZw #Innovation #Fintech #Banking #FinancialServices #Payments #Blockchain #Crypto #Stablecoin #DigitalAssets #DeFi #Infrastructure

Nicolas PintoAug 28, 2026

Payment Orchestration: Turning Declines into Routing Decisions 💡 One of the most interesting things about payment infrastructure is that a decline doesn’t necessarily mean a payment was bad. Sometimes, it simply means the payment took the wrong route. This sounds like a small distinction, but it changes quite a lot when you operate at scale. A merchant can have a perfectly legitimate customer, a valid card, enough funds and a transaction that looks completely normal and still get a decline because of the specific issuer-acquirer relationship involved in that transaction 💳 Send the same transaction through another acquirer, in the same market, and the outcome can be different. This is where payment orchestration becomes much more than having several PSPs connected to your stack 👨‍💻 The real value is not in having five providers instead of one. It is in having a system that can decide which provider should handle a particular payment, based on what it knows about that transaction and what is happening across the network at that exact moment. A card issued in the US might have a better approval probability through one provider. A European transaction might perform better through another. One acquirer might suddenly be unavailable 🛑 Another might have stronger performance for a specific combination of card, issuer, country or transaction type. Instead of sending everything down the same pipe, orchestration allows the payment infrastructure to adapt. And that logic doesn't stop at routing 🌐 If a provider declines a transaction or becomes unavailable, another route can be used. 3DS can be introduced when the risk profile actually justifies the additional friction. Volumes can be distributed across providers to avoid excessive dependency on a single partner. Over time, the orchestration layer also becomes a place where the merchant can make sense of what is happening across its entire payment ecosystem: approval rates, decline reasons, fees, chargebacks, refunds and reconciliation. That’s why I think the interesting question for payment teams is slowly changing. It used to be: “Which PSP should we choose?” Increasingly, it becomes: “How should we decide where each payment goes?” Because once you look at payments this way, a decline is no longer necessarily the end of the transaction. It can simply be a signal that the current route wasn't the best one. And if you can change the route intelligently, you may not need more customers to generate more successful payments 🚀 You may simply need to stop losing the ones you already have. Source: Solidgate - https://lnkd.in/epgSHjqv #Innovation #Fintech #Banking #Merchants #FinancialServices #Cards #Payments #Issuing #Acquiring #Orchestration #Routing