Latest / The Payments Shed / Ep. 16 - AI vs Financial Crime: Redefining Compliance, with Brian Gilman, Chief Marketing Officer, ThetaRay @ Money 20/20 USA 2025

Transcript

Ep. 16 - AI vs Financial Crime: Redefining Compliance, with Brian Gilman, Chief Marketing Officer, ThetaRay @ Money 20/20 USA 2025 · The Payments Shed

• 0:05
  Welcome back to the Payment Shed podcast walk and talk series.
• 0:07
  This morning we are joined by Brian Gilman, CMO at Feta Ray.
• 0:12
  Brian, please tell us a bit about yourself and your role at
• 0:14
  Feta Ray. Sure.
• 0:15
  I'm chief marketing officer joined in February.
• 0:18
  Beta Ray is AI powered financial crime compliance platform.
• 0:21
  Brilliant. And we wanted to talk today
• 0:23
  about that AI versus AML clash, if you like.
• 0:27
  There's a real balance that's needed.
• 0:29
  What are your thoughts with with what's happening in that area at
• 0:31
  this point of? Time I think we're at an
• 0:33
  interesting inflection point in the industry today where we're
• 0:36
  seeing legacy based rules based systems starting to really show
• 0:40
  their their cracks when it comes to a very highly regulated
• 0:43
  industry. And AI is really starting to
• 0:45
  fill that void for proving better accountability,
• 0:49
  explainability, defensibility of the data that's coming in and
• 0:53
  especially in a zero in a, in a 0 tolerance environment that
• 0:57
  we're seeing in finance. OK, perfect.
• 0:59
  And I guess it can be a really, you know, great solution to have
• 1:03
  in your, your wheelhouse AI, but it also comes with some risks.
• 1:07
  How do you look at the case study with Santander and the
• 1:10
  human trafficking example that they've they've rolled out with
• 1:13
  Federate? Yeah.
• 1:14
  So Santander is one of our largest customers and they were
• 1:17
  used there. They are using our AI solutions
• 1:20
  to really look at subtle changes in the transactions that they
• 1:24
  have in their platform. Santander Global Bank, they have
• 1:26
  millions of transactions that come to them every year,
• 1:29
  cross-border transactions and they're using our AI to look at
• 1:32
  subtlety and changes in how people are are are using their
• 1:35
  system in a rules based environment.
• 1:38
  A lot of what they were able to find with our platform, they
• 1:41
  wouldn't find because as we said earlier, as regulations changes,
• 1:46
  as compliance changes, the speed of crime is outpacing the speed
• 1:49
  of the of the legacy technology. And we're able to find subtlety
• 1:53
  and nuance in how the criminals were leveraging their platform
• 1:56
  for human trafficking, something that wouldn't have been able to
• 1:58
  be found with what they were using prior.
• 2:00
  And. What's the tangible impact of
• 2:02
  that in terms of individuals that might have been affected
• 2:05
  that you've you've potentially prevented?
• 2:06
  They were, they were able to shut down one of the largest
• 2:09
  human trafficking rings in Europe.
• 2:10
  And you know, when you think about human trafficking, drug
• 2:13
  trafficking, terrorist trafficking, there's billions of
• 2:17
  dollars flowing through the system today.
• 2:19
  And you know, equally as important on the flip side of
• 2:21
  it, because I think we talked a lot about finding the two or
• 2:24
  three bad eggs out of 100 transactions, the ability to
• 2:28
  more easily allow for the good actors to flow through the
• 2:31
  system. Is is is the other side of this
• 2:34
  which I which I don't think we talked enough about.
• 2:36
  The compliance bar is always raising, right?
• 2:39
  We see it year and year. Do you think banks are doing
• 2:42
  enough to keep up? And how can a partnership from a
• 2:45
  bank with with Federer, you know, how do they go about
• 2:48
  solving those challenges? Yeah, I think it's about
• 2:50
  accountability today. You know, when, when you look at
• 2:52
  the compliant legacy systems that are in the market today,
• 2:56
  while compliant, they garnered $4.5 billion in fines last year.
• 3:01
  And so I think we're at a point in time and you're hearing it
• 3:03
  throughout the show about I think AI is becoming more of the
• 3:07
  Norman banks are starting to figure out how do we do this?
• 3:10
  And it may not necessarily be a RIP and replace, but how do we
• 3:13
  lever an AI platform to come over the top, to prove the
• 3:16
  model, to prove the better efficiency and then over time
• 3:19
  begin to migrate? And it's easy just saying we're
• 3:21
  going to do AI or we're going to take AI, right?
• 3:24
  But it's a bit wishy washy is a statement.
• 3:26
  And I've, you know, we've had quite a lot of conversations in
• 3:27
  the last a couple of days, but I'm very interested with the
• 3:31
  bank side of things. Are you talking about as well,
• 3:33
  you know, managing these huge compliance teams or helping them
• 3:37
  to scale back those teams through the use of AI?
• 3:40
  Or is it an enabler for those teams?
• 3:43
  It's both. So if you think about a very
• 3:46
  large bank, they may have millions of transactions and to
• 3:49
  handle those millions of transactions and the false
• 3:51
  positives that are coming through their network, they may
• 3:54
  not have hundreds, they may have thousands of analysts looking at
• 3:57
  all these these false positives. So some of this is efficiency
• 4:01
  within their own network. But we talk about AI as a as
• 4:05
  wishy washy or a black box. We'll talk a little bit more
• 4:08
  about our announcement, which I think is moving the market
• 4:10
  forward. But it it's providing
• 4:14
  explainability. If I have a singular alert and I
• 4:16
  give it to an analyst in Chicago and I get that same alert and I
• 4:19
  give it to an analyst in New York, they're going to look at
• 4:21
  that alert differently. And so there's there's the
• 4:24
  people think of AI as a black box or wishy washy.
• 4:27
  But AI is also providing a level of not only credibility, but
• 4:33
  repeatability in terms of how it looks at alerts and then learns
• 4:37
  over time in terms of how to evolve those things.
• 4:39
  And it can also find nuance that the rules based systems can't
• 4:43
  today. And so I look at it much
• 4:45
  different in saying wishy washy is when you say you can either
• 4:48
  go right or left, but there is no in between.
• 4:52
  That's not good for finance, that's not good for fintechs or
• 4:55
  banks. Today we have to find a way to
• 4:56
  move the ball forward. I.
• 4:57
  Think part of the problem there is maybe just how flooded the
• 5:00
  market is at the moment. We sort of open banking, you
• 5:02
  know, we've seen it with various new innovations that have come
• 5:05
  in and I think with AI it's no different.
• 5:08
  There's, there's a huge pool of providers to choose from now and
• 5:11
  that can be a little bit confusing maybe for for brands
• 5:14
  that want to use AI, they want to explore doing something with
• 5:16
  AI, but consolidating as to who they should actually be speaking
• 5:19
  to. It's still quite difficult to
• 5:20
  vote just because there's so many players, right?
• 5:22
  Yes, and I think the part of part of the challenge is that,
• 5:25
  you know, AI for a lot of companies is a black box.
• 5:28
  They don't provide the explainability, the traceability
• 5:31
  of how they're getting to the resolution that they are.
• 5:35
  I'm going to jump to the chase. We just, we just launched an
• 5:38
  announcement today with Kaufman Rosen, who's a large, large
• 5:42
  independent audit and advisory firm.
• 5:44
  They just validated our AI models for us for both
• 5:46
  transaction monitoring as well as screening, which for us now
• 5:51
  provides a better, it takes away that arm's length discussion
• 5:55
  around defensibility of the platform.
• 5:57
  We now have third party validation, which I think is a
• 6:00
  necessary step for banks and fintechs to feel more
• 6:03
  comfortable to make that leap with.
• 6:06
  OK, This isn't just this black box that we're hoping is
• 6:09
  accurate, but our models have been validated by a third party
• 6:12
  to say this is the direction and we want that accountability and
• 6:16
  defensibility and we want to be the gold standard for for that
• 6:19
  AI intelligence. That's also the amazing news,
• 6:21
  amazing, you know, announcement for you guys yesterday.
• 6:24
  What does that lead into over the next 2-3 years for the
• 6:27
  business and why do you see the market heading as well?
• 6:29
  Yeah, you know, as I walk the floor here it was.
• 6:33
  It was interesting because six months ago there was a lot of AI
• 6:42
  common speak. I think that there's a lot of
• 6:44
  sea of sameness because AI has provided, has allowed for
• 6:49
  vendors to come into the market very quickly and say a lot of
• 6:52
  the same things. We're now hearing more about AI
• 6:55
  accountability and we're hearing more about integration.
• 6:58
  You know, the one thing that I'm seeing is a massive
• 7:01
  transformation on the floor here.
• 7:02
  When you talk about tokenization, crypto, the
• 7:05
  merging of of things like AML and crypt and cyber security, AI
• 7:11
  is going to play a much larger role in all of that.
• 7:14
  In the same breath, you have to have that level of validation
• 7:18
  for your models, validation for how you're looking at
• 7:21
  transactions. And I think that's really where
• 7:23
  it's going to separate the haves from the have nots in the space
• 7:25
  I. Completely agree.
• 7:27
  So final section of our end of the year, we are all of our
• 7:30
  guests. We have a segment called the
• 7:32
  Shell for shame where you get to nominate something in.
• 7:34
  It could be payments, it could be business, it could be
• 7:36
  fintech. Some of the really grinds your
• 7:37
  gears that you'd rather see gone forever.
• 7:40
  Alert fatigue when you think about and millions of, of
• 7:45
  transactions today and when we speak with banks and fintechs,
• 7:48
  one of the primary challenges they have is the world of false
• 7:50
  positives and the time and energy and human capital expense
• 7:55
  to just sift through the thousands of false positives
• 7:58
  that are taking place. As an AI powered financial crime
• 8:02
  compliance platform, we're hoping to remove those.
• 8:04
  You know, our platform today is, you know, can remove up to 80%
• 8:07
  of false positives. It's so challenging when you're
• 8:11
  dealing with 2030% of your overhead just trying to resolve
• 8:15
  false positives. I'm hoping that in a year we're
• 8:17
  no longer having those conversations.
• 8:19
  Brilliant. Well, happy to put that on the
• 8:21
  shelf for you. And yeah, thanks for joining us.
• 8:22
  Today, thank you so much.