Oct 7, 2026 · 18 min · 7 segments
Agentic AI has the potential to reshape capital markets as autonomous systems take on increasingly complex decisions and interactions. Based on a panel discussion during this year’s Sibos event in…
Armando BenitezGuestAkber JafferGuest
Gloria LioGuestManoj BohraGuestThat's great.
I think I see a lot of our clients moving away from an AI strategy to now the next thing is the agent strategy so that it does not get into a nightmare of too many agents and have to control agents.
I mean, I was previously working for another company where I was obviously interfacing with many different financial companies.
And this is a real challenge for some of them where they have certainly 82,000 agents.
Now, if the Fed or one of the regulators are asking you what your agents are doing, it's an impossible task for you to undertake.
It's similar to if you remember the olden time where there are thousands of spreadsheets that are actually creating your financial reporting.
And we obviously want to make sure that we balance the level of agility we provide in our ecosystem so that we are contemporary to the world that we are in.
But at the same time, we have the rigor to ensure that we do not repeat the mistakes of the past and forget the learnings that we learned.
next on that because on the tech world, you've been building a lot of these things.
And so anything to build on in terms of what they should be doing now?
Yeah, so I think maybe a slightly different lens of governance, right? So I think there's how enterprises operate and the governance around the enablement of some AI technologies, right? I do think there has to be a recognition that the world is moving at a much faster pace.
And so for sure you need governance, right? Particularly the levels that were described.
But the speed at which that governance has to move now is a lot more different, I think, right? Because things are changing on a monthly basis, right? And so therefore, there has to be an agility to governance, I would say, right? And we have to recognize that.
and view them more as enabling functions, I would say in terms of introduction of technology.
The technology itself, I think a lot has been said about observability, auditability, these kind of guardrails, right? And I'll give you a practical example and maybe slightly non-financial services, less ops, right? So we all know that Salesforce released a native connector, right? And so you can go on Claude now and go on these different things and you can interrogate it.
The kind of governance questions that I've heard of around things like that is that new connector not only allows you to read the data, but it allows you to write.
So the question is, if you have an agent that allows you to write data, that's a very different proposition from reading, right? That's the kind of level of discussions you need to have because you need to make sure you understand the risks and opportunities that come with that.
And so when we think about designing technology today, certainly the way we design our technologies, we think the sweet spot is artificial intelligence plus human intelligence is where the sweet spot is today in terms of the best outcomes.
And so something I was in Singapore two weeks ago and we talked about human being in the loop, which I'm sure we've discussed.
That's great.
I think I see a lot of our clients moving away from an AI strategy to now the next thing is the agent strategy so that it does not get into a nightmare of too many agents and have to control agents.
I mean, I was previously working for another company where I was obviously interfacing with many different financial companies.
And this is a real challenge for some of them where they have certainly 82,000 agents.
Now, if the Fed or one of the regulators are asking you what your agents are doing, it's an impossible task for you to undertake.
It's similar to if you remember the olden time where there are thousands of spreadsheets that are actually creating your financial reporting.
And we obviously want to make sure that we balance the level of agility we provide in our ecosystem so that we are contemporary to the world that we are in.
But at the same time, we have the rigor to ensure that we do not repeat the mistakes of the past and forget the learnings that we learned.
next on that because on the tech world, you've been building a lot of these things.
And so anything to build on in terms of what they should be doing now?
Yeah, so I think maybe a slightly different lens of governance, right? So I think there's how enterprises operate and the governance around the enablement of some AI technologies, right? I do think there has to be a recognition that the world is moving at a much faster pace.
And so for sure you need governance, right? Particularly the levels that were described.
But the speed at which that governance has to move now is a lot more different, I think, right? Because things are changing on a monthly basis, right? And so therefore, there has to be an agility to governance, I would say, right? And we have to recognize that.
and view them more as enabling functions, I would say in terms of introduction of technology.
The technology itself, I think a lot has been said about observability, auditability, these kind of guardrails, right? And I'll give you a practical example and maybe slightly non-financial services, less ops, right? So we all know that Salesforce released a native connector, right? And so you can go on Claude now and go on these different things and you can interrogate it.
The kind of governance questions that I've heard of around things like that is that new connector not only allows you to read the data, but it allows you to write.
So the question is, if you have an agent that allows you to write data, that's a very different proposition from reading, right? That's the kind of level of discussions you need to have because you need to make sure you understand the risks and opportunities that come with that.
And so when we think about designing technology today, certainly the way we design our technologies, we think the sweet spot is artificial intelligence plus human intelligence is where the sweet spot is today in terms of the best outcomes.
And so something I was in Singapore two weeks ago and we talked about human being in the loop, which I'm sure we've discussed.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.