EXCLUSIVE INTERVIEW - Inside Feedzai’s TRUST Framework: How AI is Powering Safer Payments Worldwide

This interview was conducted in writing and has been lightly edited for clarity.

At the HumanX Conference, Feedzai made waves with the launch of two major AI-powered initiatives—Scam Alert, a generative AI tool to help consumers detect scams in real time, and The Trust Framework, a five-pillar roadmap for building responsible AI systems in financial services. Behind these launches is Anusha Parisutham, Senior Director of Product Management at Feedzai, who’s helping shape the future of fraud prevention through ethical, customer-first innovation. 

In this written Q&A, I speak with Anusha about the role of generative AI in fighting fraud, the importance of building trust into AI systems from day one, and what trends in FinTech she’s most excited about next.

1. You lead product innovation at Feedzai. Can you walk us through what your role entails and how your team balances cutting-edge technology with responsible AI principles?

"Absolutely. Like you gave a wonderful introduction—I lead the product management for our risk platform and AI at Feedzai. What I really do is enable financial institutions to fight fraud efficiently and protect their consumers by delivering continuous innovations and intelligent automations in Feedzai’s AI-powered crime management platform.

I'm also responsible, as you lightly pointed out, for ensuring that the AI we use—we do so in a responsible, secure, and scalable way.

To your second part of the question—how do my teams balance cutting-edge technology with responsible AI? I’m going to oversimplify the answer by saying: focus on impact and being value-driven. As AI becomes ubiquitous, there’s more excitement to deliver something new and cool. You really have to step back and recognize the value AI delivers for customers—what problems it solves, and how we ensure the solutions we build are done in a responsible way.

In short, it’s not building AI for the sake of AI—but building AI that drives value to our customers and detects and prevents financial crime in a responsible way."

2. One of the major announcements at HumanX was Scam Alert. Can you tell us how it works and how it empowers consumers?

"Absolutely. Scam Alert is a fraud prevention tool Feedzai launched recently at HumanX. Scam Alert takes control back to the end consumer. Feedzai is empowering consumers to take control of scam prevention through real-time and actionable advice, which can help them identify potential scams in online advertisements or suspicious messages—even before they make a potentially fraudulent transaction.

Let’s say I’m on an e-commerce site and the price looks too good to be true, or the seller looks suspicious. I can take an image of that listing, put it into Scam Alert, and it would give me red flags or recommendations based on that image. It gives me, as a consumer, additional validation before I go through with a transaction.

It also works with phishing emails or SMS messages. You take an image and scan it. That’s what we mean by empowering the consumer and breaking the scam cycle at its source.

Scam Alert is powered by generative AI and uses multimodal large language models to analyze the content and context of those images—whether it’s a product listing, a phishing email, or a suspicious SMS. This helps detect scams early and empowers users with clear, actionable feedback.

Another benefit is that financial institutions can integrate Scam Alert into their user journeys—into their banking apps—so users can check for scams within their existing banking experience. And it’s an educational tool too. It provides recommendations and red flags in human-readable ways, so it’s not only alerting users but also educating them on what to look out for."

Follow-up: I imagine this would help everyone, but especially seniors who may not be as tech-savvy or aware of the latest scams. Is that something you considered?

"Absolutely. Elderly scams are a big concern—and unfortunately, they’re on the rise. That was a huge motivator for us. We wanted to create a solution that was not only effective, but also intuitive and frictionless for users of all ages.

With Scam Alert, all a person has to do is take an image—whether it’s of a suspicious online ad or a phishing message—and the system does the work. There’s no complicated process or technical steps to follow, so it’s especially helpful for those who may feel overwhelmed by digital tools.

And I’ll say this—it's not just seniors. Even I benefit from it! I'm in this industry, and even I sometimes second-guess certain messages or listings. Scams are becoming more and more sophisticated, and there's a constant need for education and support. Scam Alert really helps close that gap."

3. Feedzai also launched the Trust Framework at HumanX. Can you walk us through what’s included and how it helps companies operationalize responsible AI?

"Absolutely. The Trust Framework helps guide companies to be responsible with their AI practices. The misconception is that incorporating responsible AI practices comes with overhead or compromises performance or user experience—but that’s not true.

The framework includes five pillars when developing AI applications: Transparency, Robustness, Unbiased, Security, and Tested.

Let me walk through each one:

· Transparency: You need explainability, no black-box models. You should know where your data is coming from, how it's transformed, and clearly state that an outcome is AI-driven.

· Robustness: Models should be reliable and resilient. We test for adversarial threats, seasonality, and changing fraud tactics. Ongoing monitoring is key.

· Unbiased: Ensure AI doesn’t discriminate in their outcomes. That starts with representative training data and includes audits to confirm fairness over time.

· Security: Data privacy and resilience to adversarial attacks must be baked in. Secure data handling is essential in mission-critical applications like financial crime.

· Tested: Just because something is cool doesn’t mean you launch it. You test for performance, latency, cost, and more—before going to market.

It’s not just a framework for show. It includes actionable guidance and evaluation checklists to help companies adopt responsible practices effectively."

4. Feedzai uses both predictive and generative AI to fight fraud. How do those innovations improve detection while ensuring a smooth customer experience?

"From a fraud detection perspective—both predictive and generative AI improve performance by stopping more fraud with better precision. And precision means fewer false positives, which allows legitimate transactions to go through without friction.

That’s where the frictionless experience comes in—less interruption for real users, faster operations for fraud analysts, and more efficient investigations thanks to human-readable explanations and better visualizations.

So AI isn’t just making fraud detection stronger—it’s making it smoother for customers and operations teams alike."

5. Looking ahead, what trends in FinTech or AI-based fraud prevention are you most excited about?

"There are so many possibilities with the pace of AI, but if I had to pick three:

Federated Learning – This allows institutions to collaborate without sharing raw data. It’s ideal for detecting fraud patterns across institutions without violating privacy.

Adaptive, Self-Learning Models – As fraud evolves, so should our models. These can continuously improve and even use synthetic data for adversarial testing.

AI-Powered Orchestration – Imagine integrating insights across an entire customer journey—transactions, devices, behaviors, and more. That kind of orchestration enhances protection at every touchpoint."

6. For someone just starting in AI—especially in FinTech or product leadership—what advice would you give for building impactful and ethical systems?

"Start with the basics. Identify the right problems to solve and the value you’re delivering to the customer. AI tools are accessible now, so experiment, ship fast, and iterate.

Differentiation will come from customer experience. So stay close to your users. And remember—trust is king. You need to build trust into every stage of the product journey—from ideation to launch. You can’t just add it in at the end."

7. And finally, when you step away from work, what are some of your favorite hobbies or activities?

"I really love hiking—though where I live, the season is only five to six months a year. Outside of that, I do yoga and a lot of reading. I’m also a garden-to-table cook. We grow our own vegetables, and I love experimenting with flavors—especially baking. I even use AI to suggest recipe combinations! I recently did a lemon cardamom combination, and it was really amazing."

Anusha, thank you so much for sharing your insights into responsible AI, building trust, product leadership, and the future of fraud prevention. It’s inspiring to see how Feedzai is helping institutions and consumers stay ahead—ethically and effectively. 

For those who would like more information about Feedzai, you can visit their website.

Editor’s Note: This written interview was conducted by Alicia Shapiro, CMO of AiNews.com. The responses have been lightly edited for clarity and readability, while preserving the speaker’s original intent and voice. Structural formatting and editorial polish were supported by ChatGPT, an AI assistant. All final editorial decisions were made by Alicia Shapiro.