
Fighting Financial Fraud with AI: Insights from Sudhir Jha at Stanford
Insights
- AI has proven its value in fraud detection due to its scalability, measurable ROI, and adaptability to evolving threats.
- The real power of AI emerges when theoretical research from academia is paired with practical, enterprise-scale implementation.
- While AI is advancing rapidly, trust and accountability remain barriers to full autonomy in high-stakes customer-facing applications.
How is AI transforming fraud detection?
Recorded at The Business and Economics of AI workshop co-hosted by Stanford University and Infosys on May 14, 2025, this insightful interview features Sudhir Jha—senior transformation leader, AI entrepreneur, and former Mastercard executive.
Drawing on more than two decades of experience at Google and Mastercard, Sudhir explores how AI is reshaping fraud detection in financial services, emphasizing:
- Why fraud is a high-impact, real-time use case ideally suited for AI
- How enterprises can rapidly scale from data to measurable ROI
- The vital role of academia-industry collaboration in accelerating AI innovation
With strategic insight and real-world examples, Sudhir makes a compelling case for responsible, high-impact AI adoption in financial services. A must-watch for financial services leaders aiming to scale trustworthy, high-impact AI solutions.
Sudhir Jha:
I have been in the technology field for almost 25 plus years, being in AI for close to 15 plus years. Kind of accidentally got into AI with Google. Spent about 10 years there. And then the last six years I have been building AI solutions for MasterCard. And just recently started working on my startup, which is also going to be an AI solution in the fraud space.
So AI actually has been used in fraud and financial services for a very long time. And it has proven its value because of the amount of work that is involved. The scale is very big. The fraud is changing all the time, it’s evolving. And therefore, AI can really, really provide value. And it’s also very easy to measure. ROI is very easy to measure because you know exactly how much fraud you’re stopping, how much approval rates you’re increasing, how the customer experience is changing. So it's a very good application for AI.
I think there are nuances of things. So we already knew that AI is evolving very quickly. I think what we really learned is how in specific areas, there is a sort of significant change in terms of software development, how some of that is evolving into productivity gain, what is happening in terms of fraud solution, how easy it is to kind of get from data to a solution very quickly, and the enterprise AI solution that we talked about. So there is nuances that are coming out which are actually really useful to then take back to your industry and apply it.
The partnership between the universities and companies have been there for a long time. And AI, again, is a very good area because there is theoretical stuff you can do, but there is a practical application of AI that only can come from a partnership with the company. So both from a Stanford perspective, working with the company and the company working with the Stanford researchers, because again, one company cannot do all the research that is needed to advance AI-like technology.
There are areas on the front end of the business. So customer service, personalization. I mean, there is some application of AI that has happened and definitely evolving. But I think that even today, I won't trust AI agent to do everything around sort of providing both information and doing transactions on behalf of me, right? Because it's a very critical piece and the impact of a wrong decision can be very big. And so there is definitely room for innovation. And it is going to be an evolving field over some time.