Sophisticated trading strategies through automation and real-time feeds

KEY TAKEAWAYS

• Agile infrastructure is the foundation for scalability: While young firms benefit from agile infrastructure, older players have a tougher task as they transition from legacy monolithic systems toward flexible API-driven, cloud-enabled platforms. Doing so ensures firms can scale operations dramatically and become more agile thanks to customized tools built for specific regions or desks.
• Real-time data is crucial: Fast, reliable data is not a luxury but the oxygen that powers modern financial systems. It means banks can rapidly calculate complex pricing models in volatile times and ensures digital payments providers can provide operational trust and stability.
• Automation should target the client journey: True automation goes beyond superficial digitalization and involves fundamentally rethinking E2E workflows.
• AI drives efficiency, but humans are irreplaceable: Although AI increasingly enables highly personalized investment advice, dramatically improves process efficiencies and is on the cusp of transforming payments via an intent-based approach, human oversight remains critical.

The Infrastructure Shift: Overcoming Legacy Systems

As financial institutions seek ways to secure a competitive advantage, three forces are reshaping the picture: cloud infrastructure, real-time intelligence, and automation. Underlying this transformation is a central imperative: a need for agile, scalable infrastructure designed to meet the demands of modern markets.

This process is easier for some. Eric Michl, Head of Global Markets Asia at BBVA, says his bank, which is more than 160 years old, is accelerating the transformation of its technology stack toward a more agile, cloud-enabled, API-driven architecture. BBVA’s approach, Michl says, is to use API-driven solutions to build a more agile model through which data can be pulled.

“By having these flexible tools where you can go hit the data via API, traders and sales can create their own suite of products,” he says. “It’s making the tools much more powerful, and we can develop them specifically for each region or even for each of the desks. That’s really the big change.”

Globally, banks are accelerating the shift toward cloud-native infrastructure. By way of example, says Michl, BBVA partnered with Bloomberg and Amazon Web Services to replace its volatility marking system. Although it took two years to build a cloud-hosted replacement, he says, “There’s no looking back – it’s clearly a game-changer that’s allowed us to scale our business, not only in Asia, but across the globe.”

BBVA’s experience compares with that of Singapore-headquartered dtcpay, which offers real-time multi-currency swaps between stablecoins and fiat currency. Hong Kong CEO Candice Zhang says dtcpay has an advantage as a six-year-old digitally native firm, which it used to best effect early on. Knowing the challenges faced by traditional OTC players, dtcpay invented its first OTC product – a fully automated engine for its clients.

“That was our starting point,” she says. “And we’re lucky we’re in Asia, because [in this environment] people are used to real-time payments in the traditional world. And so, when it came to stablecoin payment industry, [our product] was not a hope. It’s what they need.”

And while some challenges are beyond dtcpay’s control – cross-border payments typically settle on T+1, while traditional banks are closed after-hours and on weekends – Zhang is hopeful that real-time, 24/7 automated payments will be a feature of the region sooner rather than later.

The Nuances of Real-Time Intelligence

As Bloomberg has noted previously, high-quality, real-time data has become central to investment decision-making and risk management. However, when it comes to real-time data, different players have different needs. That reflects the fact that real-time is not a monolithic concept across the financial spectrum. While data velocity remains vital, its application and the value financial institutions extract from it often diverges sharply.

For dtcpay, operating as it does in a fully regulated and fast-moving environment, real-time data is essential for risk management. Data, as Zhang puts it, “is our oxygen: If there is no real-time data, no reputable, reliable data, it’s just nothing – because trust is our currency.”

Alex Ng, Head of Investment Services, says that need is less marked for UOB’s private banking clients.

For retail investors with medium- to long-term horizons, “real-time” does not necessarily mean immediate, millisecond-level streaming. Instead, real-time data is about ensuring the “correctness and transparent mark-to-market” of a client’s multi-asset portfolio, which could contain a range of assets including equities, FX, structured products and derivatives, for which price feeds vary.

“So, what’s important for us is actually the client statement and client presentation – so that at the point in time that the client sees his portfolio it’s accurate and it’s timely,” Ng says. “Timely here doesn’t mean immediate; it has to be relevant. So, for us, real-time data reflects relevance and trust.”

Automation Driven by the Client Journey

As financial institutions continue trying to extend their competitive advantage, it is increasingly the case that automation needs to transcend superficial, surface-level digitization. Today, building a sustainably different approach comes not through isolated technological upgrades, but via successfully integrating advanced capabilities into a seamless, cohesive E2E workflow.

For UOB’s Ng, successful automation requires targeting the client experience. Where previously best execution for private banking covered simple venue-driven pricing, today it means managing the E2E client journey. That encompasses front-end processes like client onboarding, KYC, and risk suitability checks, all the way through order placement and post-trade performance measurements.

“That entire client journey for us, end-to-end, is really the key to delivering the trust that the client is buying, procuring the right investment products that are suitable for the client to express their investment objectives,” he says.

The importance of viewing automation through that client-service lens was emphasized in a later discussion between Jeff Hutchins, Head of Equities Japan and Head of Asia Equity Risk Trading at Jefferies, and Bloomberg’s Heena Parekh, Head of Enterprise Data & Technology Sales, North Asia.

As Hutchins explains, Jefferies operates as an advisory-driven firm, generating around 90 percent of global revenue from client fees for advice and trade execution. Given that its success relies on clients actively valuing and paying for its services, Jefferies’ strategy is rooted in directly addressing client needs by asking them directly, rather than assuming what solutions might be a useful fit.

To that end, Hutchins says, Jefferies does not pursue automation for the sake of modernizing. And, he notes, clients rarely ask for automation. They want tangible, practical improvements like lower risk, fewer errors and faster time-to-market.

“Automation is not the goal. The goal is solving problems,” he says. “Now, automation is often a mechanism to solve problems in a very good way, because with automation, risk can go down … throughput can go up, things can be scalable, you can trade more things with less people, with less errors. We don’t just automate for automation’s sake; we automate to solve a problem for a client.”

The Future: AI and the Human Connection

Asked for their views on developments over the coming 18 months, BBVA’s Michl predicts shorter settlement cycles, longer trading hours, and AI-driven personalized advice based on deep integration with client portfolios. For dtcpay’s Zhang, automation will soon move further up the funnel to “intent-based payment,” where an AI agent simply finds the best pricing and executes a transaction based on a user’s verbal command to their phone.

Finally, says UOB’s Ng, although AI can transform process efficiency, its application is less clear in more nuanced areas like determining a client’s risk appetite. The contradictory nature of human behavior means a client’s questionnaire might yield a very conservative risk profile, even as their portfolio is filled with high-risk investments. As algorithmic models struggle to reconcile such disconnects, human intervention is crucial.

“Even if you did put the right data into some agentic AI to generate a certain kind of risk profile, when you go back interviewing the client … emotions take over,” Ng says. “He would tell you that ‘No, I don’t like to lose money, but I like to invest in the risky stuff.’ So, you have that disconnect … which is why that human element is still quite critical in what we do.”