Скачать или смотреть Panel Discussion: Minimizing Fraud Risk in Financial Transactions with Hyperautomation

Panel discussion on “Minimizing Fraud Risk in Financial Transactions with Hyperautomation,” featuring industry-leading experts from Wells Fargo, AWS, Capgemini, and Waylay.
Industry panelists include:

Ajit Gaddam, SVP, Head of Fraud Technology, Wells Fargo
Paul Chang, Head of Payment Networks, AWS Financial Services
Rajesh Iyer, VP Global Head of AI/ML, Capgemini
Shankar Ramanathan, Global AI/ML Leader, Capgemini
Veselin Pizurica, COO & Co-founder, Waylay
Kris Vandenberk, Director Technology, Agio Capital

0:00 Kris Vandenberk: Introduction
1:12 Ajit Gaddam: shares some of the most notable trends and developments in financial fraud, as financial institutions and regulated industries face ever more complex money laundering and criminal schemes.
2:32 Ajit Gaddam: how do we move from a rule-based or simple automation ?
3:01 Paul Chang: perspective from Amazon's point of view
3:40 Paul Chang: Juniper research forecasts online payment fraud will exceed 343 billion for the period between 2023 and 2027.
3:48 Paul Chang: we see banks and payment companies continue to invest in security technologies
6:01 Shankar Ramanathan: Capgemini view on perspectives and trends. Customers right now are straddled between two worlds, the original world where the number of attributes that are coming in number of features that are coming - it is becoming more and more unmanageable
10:27 Rajesh Iyer: Introducing the concept of hyper automation in terms of evolution. Many of us can think of this as RPA but that is not really and answer as RPA can’t actually "stitch" these things together at the required speed
13:48 Veselin Pizurica: RPAs are effective in places like document processing or market automation but specifically when we talk about finances that orchestration technology really has to deal with information with unprecedented speed and scale.
14:58 Veselin Pizurica: In the past people talked about rules-based systems as the means to express knowledge. Over the time people have started thinking about ML models as a means to delegate decisions, what are we actually observing in many cases that it's not one or the other but basically you have to use best of both worlds
17:19 Paul Chang challenges that we hear from customers to have good fraud detection. An essential tool for detecting and preventing fraud are cloud platforms that provide the scalability and compute power
18:01 Paul Chang banks have historically been more batch based, they're starting to leverage definitely real-time data and the payment world is going to move towards more data consortiums.
19:17 Ajit Gaddam: the evolution towards a Consortium and data sharing.
21:28 Paul Chang: Trends that's going to push more towards data sharing on is real-time payments right as banks adapt to ISO 2022 standards
22:19 Veselin Pizurica: I see the evolution from the stream processing engines to more flexible real time engines.
23:04 Shankar Ramanathan on how to help organisations from the legacy state to a new era of new technology
24:04 Shankar Ramanathan a new hybrid way of handling things with ML and policies, A/B testing
26:20 Rajesh Iyer: when building a solution you're looking for four things to get better
28:23 Ajit Gaddam from a bank or a financial institution standpoint, how we can react faster
31:15 Veselin Pizurica: How to benefit from the latest technology advances in cloud computing, and shorten the adoption cycle that is not only about cloudification but to truly benefit
from cloud native approach.
33:52 question from from the audience, Kumar, how hyper automation can help with identity theft of or data overload attacks?
37:33 Paul Chang talking about ML providers and how these information is helping orchestrate these outputs and checks so you have a company like Waylay that's working with multiple vendors orchestrating these capabilities and helping banks and fintechs
38:23 Paul Chang on vendors that are very mature in the US as they have an entire Global database that has captured photo IDs with fraud
39:05 Shankar Ramanathan on synthetic ID and device profile check, and compromised email addresses
39:44 Shankar Ramanathan on a need for flexible and conditional real time checks
39:58 Rajesh Iyer important factor in how these applications need to work together using low code and orchestration paradigm
41:31 Rajesh Iyer on importance to start with simple things first
42:43 Veselin Pizurica “stitching things together” it's a very complicated and very complex technological challenge
43:46 Kris Vandenberk: is low code a gimmick or a real deal for high transaction volumes?
46:09 Ajit Gaddam: How do we account for data governance?
50:00 Compliance and explainability
52:06 Do we foresee fraud decisioning moving more towards AI based solutions?
55:38 Final thoughts