Latest / The Payments Shed / Ep. 15 - AI Coworkers for Fraud Teams, with Riya Jagetia, Co-Founder @ Socratix AI @ Money 20/20 USA 2025

Transcript

Ep. 15 - AI Coworkers for Fraud Teams, with Riya Jagetia, Co-Founder @ Socratix AI @ Money 20/20 USA 2025 · The Payments Shed

• 0:07
  Welcome back to the Paper Shed podcast.
• 0:09
  Today we are joined by Rio Jagatia, CEO and cofounder of
• 0:12
  Socratics AI. Tell us a bit more about
• 0:15
  yourself. Yeah, my name is Ria.
• 0:17
  I am the cofounder and CEO of Socratics.
• 0:20
  We started about six months ago, but I've previously been in the
• 0:22
  fraud space for the past four years, building solutions at
• 0:25
  DoorDash and before that on the vendor side at Unit 21.
• 0:29
  I realized that fraud teams deal with a lot of manual reviews, a
• 0:32
  lot of manual work flows, and wanted to build an AI solution
• 0:35
  that helps them basically lighten the load of what those
• 0:38
  teams have to have to deal with. Fantastic.
• 0:41
  And the industry, as we know here in Money 2020 as well is,
• 0:44
  is full of tools that that use AI to assist teams, right.
• 0:48
  What does AI mean to you when it comes to instead of kind of
• 0:53
  creating dashboards after dashboard, but really helping
• 0:56
  businesses not have to go through that manual labour?
• 0:58
  Yeah, no, it's a great question. I think that's why we really
• 1:00
  frame our solution as AI Co workers, because we don't just
• 1:03
  want another tool that people have to take a look at and you
• 1:06
  know, get information from and then action on.
• 1:08
  We really want to build something that can work to our
• 1:10
  customers, action for them, bring them new insights, stitch
• 1:14
  together information from different sources, really acting
• 1:17
  like an assistant rather than just another tool that you have
• 1:20
  to check everyday. I think we, we see it quite
• 1:22
  often, right? Businesses, fintechs and banks
• 1:25
  roll out what they think is a new product internally.
• 1:28
  It takes 2 years for them to really roll out.
• 1:30
  And by the time it's live, it's already behind the times.
• 1:33
  Yeah. Where do you think you guys with
• 1:35
  AI? I'm going to assume that it's
• 1:37
  updating all the time and always getting more intelligent.
• 1:40
  Yeah, yeah, that's a great point like that.
• 1:42
  I think that's one of the ways that our solution is different
• 1:45
  from traditional dashboards or tools.
• 1:46
  It's constantly evolving, constantly changing and updating
• 1:49
  based on human input, but also new data.
• 1:51
  So suppose you bring in a new vendor and that vendor brings
• 1:54
  you more data that's useful. You would usually have to update
• 1:56
  a dashboard, update your logic, your rules to make sure it's
• 1:59
  taking that into account. Our AI agents can just learn
• 2:02
  from what your analysts are doing to make sure it's acting
• 2:04
  the same way. Makes completely sensitive when
• 2:08
  we go and speak to most businesses that I have when when
• 2:11
  trying to put new products and roll out new products across
• 2:13
  their teams. The issue we normally have is
• 2:16
  the challenges that come with it and the dread right, of having
• 2:19
  to roll a brand new product. What are your thoughts on on the
• 2:23
  way that the AI is going to roll out compared to how it is maybe
• 2:26
  being rolled out in fintechs and banks today?
• 2:28
  Yeah. No, it's a great question.
• 2:29
  I think that was one of the things we wanted to make sure
• 2:31
  was a major differentiator of what we're building versus what,
• 2:35
  you know, potentially legacy customers are asking banks and
• 2:38
  fintechs to do. So we don't say send us all of
• 2:40
  your data onto our platform, do a really long integration and
• 2:43
  then we have all the bells and whistles, bells and whistles on
• 2:46
  top. Instead we plug into customers
• 2:49
  existing systems. So we have a proprietary kind of
• 2:51
  infrastructure that we've built that allows us to get customers
• 2:54
  up and running in a matter of days.
• 2:56
  We have kind of custom connectors and then we just work
• 2:58
  with their existing tools to complete workflows the way that
• 3:00
  a person would. Yeah.
• 3:02
  And you mentioned prior to us recording, yeah around how
• 3:05
  quickly it quick it is for businesses to to get live and
• 3:08
  running with you. Yeah.
• 3:09
  I assume that's going to be a big differentiator for you as a
• 3:11
  business, right, Because you're doing a lot of the heavy lifting
• 3:13
  on their. Behalf exactly.
• 3:15
  Yeah, it's a huge differentiator because like I said, our first
• 3:17
  customer, customer was live within three days.
• 3:20
  Most of our customers are live within a week at most.
• 3:23
  And that makes a big difference because they can test out the
• 3:25
  solution, look at the results before even entering into kind
• 3:29
  of a long term agreement, and that makes a world of
• 3:31
  difference. I bet, I bet.
• 3:32
  So we talk about AI investigating fraud and I think
• 3:35
  here at Money 2020, I probably spoke to three or four
• 3:38
  businesses that are changing the world, right, When it when it
• 3:41
  comes to AI, Yeah, it's risky, right?
• 3:43
  Yeah. And we all know that.
• 3:45
  How do you think you provide comfort to the risk teams and
• 3:48
  maybe people that are seeing sometimes a glass is half are
• 3:51
  half empty as opposed to a glass half flow?
• 3:53
  Yeah, yeah, there's, there's basically two things that we do.
• 3:55
  So the first is we allow them to have human in the loop controls
• 3:58
  for any high risk workflows. They could do it for all of
• 4:01
  their workflows. So customers usually do is for
• 4:03
  the first couple of weeks, we'll have humans review every AI
• 4:06
  agent output before it takes an action.
• 4:09
  And then they usually say, can you please have the AI agents
• 4:12
  automatically start to take action because the accuracy is
• 4:14
  really high. And at that point, we say, yes,
• 4:17
  we can definitely take action automatically for you.
• 4:19
  But just like you would for a person, we want you to have QA,
• 4:21
  QC. So maybe 10% of the AI agent
• 4:24
  reviews every week someone else takes a second look at just to
• 4:27
  make sure that as your patterns changes, as your risk appetite
• 4:30
  changes, our AI agent is reacting appropriately.
• 4:33
  Makes sense. Thank you very much.
• 4:34
  So yeah, what are your thoughts so far in Money 2020?
• 4:37
  We know it's an incredibly about 13,000 people.
• 4:39
  Yeah, it's a lot to take in, right?
• 4:40
  How have you found it so far? It's been amazing.
• 4:43
  I feel like AI is the name of the game everywhere.
• 4:45
  AI and probably stablecoins, yeah.
• 4:47
  And so I think what's been really interesting is trying to
• 4:49
  dig a step deeper or a level deeper with customers to figure
• 4:53
  out what do they actually mean when they say AI, How is it
• 4:56
  actually being used? What are the measurable results?
• 4:58
  Because I think ROI is so important when you're deploying
• 5:01
  new solutions. And yeah, it's been fun to see,
• 5:04
  you know, all the information that customers here or companies
• 5:07
  here have to offer. Fantastic.
• 5:09
  And last thing here on the shelf of shame, Yeah, we love it here
• 5:12
  at the Payment Show podcast talking about really what grinds
• 5:15
  your gears. Tell me a little bit more about
• 5:16
  what really gets to you here in the paying some fintech
• 5:19
  industry? OK, so I think I'm, I'm really
• 5:20
  in the fraud space with like the fraud and compliance space.
• 5:23
  And there's this term that was coined maybe 5 or 6 years ago
• 5:26
  called framel. It's fraud and AML, and that
• 5:29
  really grinds my because I think it's so important when you're
• 5:31
  developing products to have a really clear definition of what
• 5:34
  every team does and what they care about and what their
• 5:38
  motivations are. And fraud and risk and
• 5:40
  compliance actually are three completely kind of separate work
• 5:43
  flows with separate motivations. So it really grinds my gears and
• 5:46
  people try to combine them all into one because then you're
• 5:48
  just not understanding the customer super well.
• 5:50
  So, you know, if I could change the world, I would say let's
• 5:52
  let's start talking about these three teams separately, even if
• 5:55
  they use the same software. So we deeply understand their
• 5:57
  needs, their concerns and what to build for them.
• 5:59
  Fantastic, so we need to retire the word framel.
• 6:02
  Retire that, just say fraud risk compliance or AML and be really
• 6:06
  clear about what you're solving for them and why and what their
• 6:08
  return is. Fantastic, thank you so much for
• 6:10
  your time. Enjoy the rest of the show.
• 6:11
  Thank you SO. Much.