FNG Exclusive Interview… As trading platforms race to differentiate in an increasingly competitive market, Deriv is betting that AI-powered intelligence will be the key to faster innovation and better client outcomes. FNG spoke with Xavier Gorriz Torner, Deriv’s VP for Business Intelligence and Data Engineering, about how the company is rebuilding its data operations from the ground up and what that means for traders.

FNG: Deriv has been vocal about becoming an AI-first organisation. What role does data infrastructure play in that vision?

Xavier: Data infrastructure is the foundation that makes everything else possible. You can’t automate customer service, payments, compliance, or marketing without a data layer that can instantly access, process, and analyse information across all systems.

Our vision is to automate everything that can be automated. But automation at scale requires intelligence at scale. Over the past ten months, we’ve rebuilt our data architecture from the ground up. Our data layer now processes over 1 million API requests monthly, a tenfold increase. Internal queries that took days are now complete in seconds.

That speed and reliability enable Amy, our customer service AI, to handle complex queries autonomously. It allows payment systems to process transactions at maximum speed. It makes automated compliance workflows possible. Without the data foundation, none of the automation works.
For us, being AI-first means fundamentally rethinking how we work with data. Instead of bolting AI features onto existing processes, we’re rebuilding our infrastructure from the ground up so intelligence is baked in from the start.

But here’s where it gets interesting for traders. The same automation philosophy we’re applying internally will eventually transform how clients interact with our platform. Imagine tools that automatically analyse trading patterns, surface insights that might have been missed, and help make more informed decisions without having to dig through data manually.

This isn’t a pilot programme. This is how Deriv operates now.

FNG: What does this data infrastructure actually look like?

Xavier: We’re building an ecosystem of specialised AI agents that work together autonomously. Teams can now request data in plain English. Someone types “show me verification completion rates in Malaysia last quarter” and the system understands the query, accesses the relevant databases, and returns results instantly.

We have agents that automatically document datasets, design data structures, and maintain data quality. When new systems come online, agents map the data relationships and make that information immediately accessible.

The infrastructure also supports proactive intelligence. We’re developing agents that continuously monitor metrics, detect anomalies, investigate root causes, and alert relevant teams. Imagine withdrawal processing slows at 3 AM. The system detects it, traces it to a payment provider issue, and alerts the team with a fix before Asia’s morning trading session.

FNG: How does this create competitive advantage in the brokerage space?

Xavier: Speed of execution. We can analyse millions of verification data points instantly, identify friction points in real time, and push fixes into production the same day. Traditionally, brokers need analysts to write queries, wait for results, present findings, and then wait for decisions.

Content creation is another example. We’re automating educational materials, ebooks, videos, and marketing assets across all languages. That requires data infrastructure that understands client behaviour, identifies knowledge gaps, and feeds intelligence to content generation systems. What took weeks now happens in hours.

Our AI systems design PPC campaigns optimised for ROI, interact with partners to maximise performance, and adjust spending by country based on local regulations and competitor activity. This level of optimisation at scale isn’t possible with manual management.

FNG: What about data quality and security at this scale?

Xavier: Security and compliance are non-negotiable. Every data access requires strict controls and audit logs. We’ve built validation layers throughout the system because AI can occasionally generate plausible but incorrect information. We catch these through automated verification that cross-references multiple data sources and flags inconsistencies.

We have AI systems monitoring for intrusion attempts in real time. Any unauthorised access gets detected and blocked immediately. For compliance, AI generates regulatory forms and license applications ready for human review. Client assessments and KYC processes are fully automated with audit trails at every step. In financial services, you can’t compromise on data integrity.

FNG: What are some obstacles in building this?

Xavier: The biggest is maintaining 100% uptime while rebuilding infrastructure at full capacity. We can’t afford downtime for a global trading platform. We’re upgrading our entire data architecture whilst keeping everything running smoothly, like rebuilding an aeroplane mid-flight. AI accuracy at scale remains challenging; with large data volumes, AI can sometimes generate plausible but incorrect information. We’re addressing this, but at a massive scale, it’s still an industry-wide challenge.

Data sprawl is another challenge. Information is stored across dozens of systems with varying structures and formats. Teaching AI agents to understand these relationships and query efficiently requires extensive mapping work.

FNG: How does this translate to business results?

Xavier: We’re launching products and features in weeks that used to take quarters. When your data infrastructure can answer complex questions instantly and your compliance workflows run autonomously, you remove the bottlenecks that slow everything down.

We’re also seeing it in talent retention. Engineers and analysts don’t want to spend their careers running manual reports. They want to solve interesting problems. When we automate repetitive work, we keep our best people engaged and productive.

The market will ultimately judge brokers on client outcomes. Faster onboarding, more responsive support, better educational tools, and trading platforms that adapt to individual needs, all powered by the underlying data infrastructure. This is where AI becomes transformational for trading platforms. Not just automating back-office tasks, but fundamentally changing how traders interact with markets. All of it depends on having data infrastructure that can process intelligence at scale. That’s where this investment pays off.