Ecouter sur Spotify 
Ecouter sur Apple podcast 
Ecouter sur Deezer 
Ecouter sur Ausha “On cybersecurity, it’s the first time in a decade, I’d say, that the cost of defending is higher than the cost of attacking. Five years ago, hackers were spending more money trying to hack into systems. And of course, they were trying to access very sensitive data, to sell it on the dark web, and so on. Today, all the analysis shows that it’s the opposite: we need to spend and invest more money than they do. And AI is helping with cybersecurity, and should be helping even more going forward.”
Eric Bierry, CEO of SBS & Deputy CEO of 74Software, speaking on the What Leaders Want podcast
AI adoption in banking has moved from “if” to “how”
A year ago, many banking leaders were still asking whether they should adopt AI at all. According to Eric Bierry, CEO of SBS and Deputy CEO of Exway, that debate is over. The question banks are asking now isn’t whether to use AI, but how to use it, for what purpose, and with what outcome in mind.
Eric Bierry points out that AI’s core techniques aren’t new: many of the algorithms behind today’s tools were already understood 10 to 15 years ago. What changed is the computing power available to run them and the speed at which they’ve spread. Where technologies like the internet or mobile took several years to reach mass adoption, Bierry says AI has produced comparable disruption in banking in roughly a single year.
How AI is changing the day-to-day work of bankers
For client-facing staff, AI is enabling what Eric Bierry calls “augmentation”: bankers can prepare a full 360-degree view of a client in minutes instead of hours, anticipating needs (a new home, a growing family) before a meeting even happens. He expects banks to employ fewer relationship managers over time, but ones who are better equipped and have more time to focus on the client relationship itself, rather than routine research and prep work.
Behind the scenes, AI is also helping teams identify unusual behavior and correlate larger volumes of data to flag potential fraud faster. Just as importantly, it’s cutting down false positives ; a change that frees up banker time previously spent confirming that flagged transactions were legitimate.
Cybersecurity: the economics have flipped
One of the more striking points Bierry raises is around cybersecurity spending. Five years ago, attackers generally spent more than defenders to break into systems. Today, he says, that balance has reversed: banks now need to invest more in defense than attackers spend trying to breach them ; a direct consequence of AI-driven threats and the growing volume of real-time data that needs protecting.
Why “human in the loop” isn’t going away
Despite the push toward automation, Bierry draws a clear line: clients should always give explicit approval before a transaction executes, whether that approval happens in real time or is anticipated for a recurring payment. He doesn’t expect that to change in the next decade, even as more of the surrounding process (options, recommendations, anticipation of client needs) becomes automated.
What this means for banks and their clients
For clients, the near-term impact is more options presented earlier in a conversation, as relationship managers use AI to prepare beyond what they could anticipate alone. For the industry, it means faster product cycles (a two-year delay that was tolerable a few years ago is no longer acceptable) and a need to stay agile as new models continue to make previous ones obsolete within months.
This is Part 1 of a three-part conversation with Eric Bierry. Part 2 (coming next week) covers why SBS is choosing a different path than the major US cloud platforms for running its AI.
