In a world where markets can crash with a tweet and regulatory shifts strike overnight, financial institutions must react in milliseconds. Traditional data systems are too slow. That’s where real-time data streaming comes in. We spoke with Niki Hubaut, Country Leader France at Confluent, to understand how streaming architecture is helping trading floors stay resilient and keep a strategic edge—especially during times of crisis.
1/ What is data streaming, and how does it enable faster reactions on financial markets?
Niki Hubaut: “Data streaming consists of processing data in real time, at the moment it is generated, instead of periodically in batch. Confluent is a real-time data streaming platform built on the open-source solution Apache Kafka, enriched with features that enable the processing, governance, and distribution of event-driven data. It captures every event (e.g. market order, user click…), transforms it on the fly, and routes it to the relevant systems (analytics databases…) with extremely low latency, in the order of milliseconds.
Data streaming is an advantage on financial markets because it allows:
- To reduce decision latency (react to the event as soon as it’s generated)
- To have a unified real-time view (all operational flows—transactions, orders, positions, risks—are normalized and can be continuously correlated)
- To have stream-native computation engines (complex computation pipelines coded directly on the flows, e.g. risk scoring or abnormal behavior detection)
- To automate reactions (an alert, an order block, or a notification can be triggered automatically without human intervention)
On financial markets, everything can be decided in milliseconds. With data streaming, we can detect a market trend live and adjust positions immediately, spot fraud or suspicious transactions as they occur, react to a regulatory or rate change in real time, and continuously feed trading algorithms or risk models without delay.
Data streaming allows us to act on data the moment it passes. In finance, that means always staying one step ahead of the market.”
2/ People say real-time data is the new oil. Is that an exaggeration in the context of financial markets?
Niki Hubaut: “Because markets have become instantaneous, simply owning data is no longer enough. It must be captured, analyzed, and—most importantly—acted on before others do.
That’s exactly what data streaming allows: processing data as soon as it arrives, within a few milliseconds, and triggering business decisions without waiting for an hour, a day, or a batch at the end of the day. In a market where every millisecond counts, this is simply a strategic advantage.
For financial players, real-time data allows them to maximize profits, reduce risks, maintain compliance, and optimize the customer experience. What Confluent offers is a platform where data is captured in real time (Kafka), intelligently transformed (Flink, ksqlDB), distributed where it has value (trading system, dashboard, AI) and governed securely (Schema Registry, access control, audit).”
3/ What are the most common blind spots among banks or trading platforms when it comes to data flows?
Niki Hubaut: “The most frequent blind spots we see are:
- Hidden latency in the architecture (many think they are “real-time,” but in reality their pipelines introduce latencies of several seconds or even minutes)
- Unconnected application silos (data flows are often fragmented between teams or systems: trading, risk, compliance, IT, data science… each with their own tools and data access)
- Lack of traceability and governance in streaming (e.g. versioning and event replayability)
- Gap between real-time ingestion and business usage because data remains unusable without prior storage
- Security poorly adapted to the event-driven model (exposing sensitive events to unauthorized users)”
4/ Can data streaming really help avoid irrational decisions during financial crises like the one we’re seeing today?
Niki Hubaut: “Data streaming allows for immediate access to complete information, reactions based on rules or models, continuous updates of key indicators (decisions are based on the system’s actual state, not on an outdated snapshot) and alignment between teams (access to a single source, reducing internal friction and contradictory decisions).”
5/ What role does artificial intelligence play in this equation? Can it be effective without real-time data?
Niki Hubaut: “AI makes it possible to anticipate market movements (prediction), detect abnormal behavior (fraud, manipulation), dynamically adjust trading strategies, personalize the user experience based on context, and help manage risk in uncertain environments. But all this depends on having reliable and up-to-date data. Without real-time data, AI learns or reacts to already outdated scenarios. It misses weak signals. Also, it cannot act within the decision system (which moves very fast). And it becomes a slow analytical machine, useful after the fact but inoperative in the moment.”
6/ And after streaming—what do you see as the next big data revolution in finance?
Niki Hubaut: “The next revolution will be the fusion of data streaming, AI, and automated actions. We’re moving toward systems that not only react in real time but also decide, prioritize, and orchestrate complex responses without human intervention. For example: automate business decisions based on real-time data, AI models, business rules, and continuous feedback; stop waiting for user actions and start suggesting actions based on context. It’s about enabling data to act intelligently within a controlled framework.”
Interested in knowing more about how AI is reinventing financial services? Read our story:
