AI's Real Value Isn't the Glamorous Stuff: Euroclear on Data Quality

Quick Summary

With AI the headline theme of Sibos 2026, Isabelle Delorme, Global Head of Product Strategy & Innovation at Euroclear, argues that the shift has already happened. The real question for a market infrastructure sitting on massive volumes of data is how to use AI in depth, and that starts with data quality and data governance. The most exciting initiatives, she says, are the least glamorous: operations, efficiency and better controls.

Is AI still "the shift" for market infrastructures?

Not according to Delorme. The shift already happened, she says. We all use AI all the time in our private lives, and it is now the case that we use it professionally too. So the conversation is no longer about whether the shift is coming.

What matters for capital markets is something more specific. Infrastructures like Euroclear sit on massive volumes of data, so the opportunity goes beyond the bread and butter capabilities of AI. It is about using AI in depth to improve the way data is managed. Delorme says a lot has already been done, but the industry is probably only at the beginning of what AI can do.

Why is data quality the big investment?

For all market infrastructures, banks and CSDs, Delorme says, a lot of the work is about data governance and data quality, because using AI on unreliable data simply would not work. She puts it bluntly: AI on data that is not totally reliable is not going to deliver.

That is why the big piece of investment is around data quality and data governance. Only once that foundation is in place will the industry be able to extract the full benefits of AI.

Where are the most exciting AI initiatives?

Delorme says it is fair to say the most exciting initiatives are maybe not the most glamorous. They are about:

• Operations and efficiency
• Using AI for better controls
• Automating processes that were not manual, but were a bit fragmented

Her message to the industry is that it should not look only at the sexy and fancy things AI can do, but at how to better manage data across the board.

FF NEWS TAKE:

This is the most useful kind of AI answer at Sibos, because it refuses the hype. A central securities depository does not need a chatbot story, it needs clean, governed data and tighter controls. Delorme's point that AI on unreliable data will not work is obvious once said, yet it is exactly where many programmes stall. If this year's theme is digital finance in an AI-driven world, the infrastructures that win will be the ones that did the unglamorous data work first.