Summarize with:
In an exclusive interaction with CXO Media & APAC Media, Koushik Kadidal, Chief Data Officer & Head of Insights Business, PayU explains how GenAI is being leveraged to improve risk assessment and fraud detection in digital payments. He also outlines how PayU ensures data privacy, security, and compliance with the rise of AI-driven personalization.
PayU has announced its strategic focus on Generative AI for 2025. Could you elaborate on the key AI-driven initiatives currently being implemented to enhance customer and employee experience?
At PayU, our GenAI implementation strategy focuses on three core areas across our businesses: enhancing merchant experiences, optimizing operational efficiencies, and empowering our employees.
For merchants, we have implemented GenAI systems that have significantly streamlined onboarding, enhanced fraud prevention capabilities, and transformed customer services. We have observed an 80% reduction in fraud and chargeback losses while accelerating the onboarding of top category merchants to just 2-3 days. Our GenAI-powered email classification system is transformative and has already helped us resolve 30-40% of inquiries instantly, dramatically improving our response times. As a result, our care agents save up to 7% of their processing time.
Internally, we have worked towards enhancing our development teams’ efficiency by deploying our in- house GenAI platform, Toqan, that facilitates code reviews, conducts security evaluations, and enables text-to-code conversion with 65% accuracy, resulting in 15-20% time efficiency. Toqan has transformed employee workflows by enabling secure, compliance-aligned access to premium AI models, revolutionizing how teams gather information, create reports, and leverage data for decision-making processes.
In addition to this, for our credit business, we have made significant strides in democratizing GenAI access. All employees have access to advanced GenAI platforms such as OpenAI, Meta, Anthropic, and others within a robust InfoSec and Compliance framework. Moreover, PayU Finance employees now have access to backend databases without requiring prior SQL knowledge, which has substantially improved data accessibility and decision-making across the organization.
How is PayU leveraging Generative AI to improve risk assessment and fraud detection in digital payments? Can you share insights into the AI models and frameworks being used?
Our GenAI approach to risk assessment and fraud detection operates on multiple levels. To share a broader sense, we have implemented sophisticated GenAI models that analyze transaction patterns, merchant behavior, and market anomalies in real-time, which have resulted in an 80% reduction in fraud and chargeback losses. These systems efficiently identify discrepancies between merchant-provided information and public data during onboarding, flagging potential issues for human review.
Our framework also combines proprietary risk engines with leading GenAI models, all within a human-in- the-loop approach that ensures accuracy and fairness. For instance, Wibmo, our enterprise Fraud Risk Management platform, utilizes advanced rule engines to detect anomalies, fraud rings, and velocity bursts
in real-time.
in real-time.
We are currently exploring the integration of GenAI to enhance this system that provides risk analysts with actionable insights that can further minimize fraud losses and foster ecosystem trust. All these capabilities operate within a robust governance structure that ensures compliance with regulatory requirements while maintaining the agility needed in fraud prevention.
With the rise of AI-driven personalization in fintech, how is PayU ensuring data privacy, security, and compliance while using Generative AI for customer engagement?
At PayU, we take a responsible approach to GenAI, ensuring that all our GenAI implementations comply with regulatory standards while prioritizing security and ethical usage. We have implemented a comprehensive framework that includes stringent access controls, data minimization principles, and continuous security monitoring.
All GenAI applications operate within our InfoSec and Compliance framework, adhering to global standards including GDPR, RBI guidelines, and PCI-DSS. For customer engagement, we employ privacy-by-design principles where GenAI models are trained on carefully curated datasets that minimize exposure of sensitive information. We have established clear audit trails and explainability mechanisms that provide transparency into how decisions are made.
Our no-learning policies for certain applications ensure customer data is not used to improve models without explicit permission. This balanced approach allows us to deliver personalized experiences while maintaining the highest standards of data protection and regulatory compliance.
What are some of the biggest technical challenges in deploying Generative AI at scale within PayU’s payment ecosystem, and how is your team addressing them?
So far, our GenAI journey has revealed two key hurdles. First, accuracy improvements with GenAI require an exponential effort. So, while the first set goal to achieve accuracy can be easy, achieving each additional improvement demands exponentially greater investment in model refinement and data quality, which adds to making the process daunting and increases lead time. To navigate this, for instance, we have only deployed GenAI for first-level responses and initial query classification, rest all follow-up communications and escalations are handled personally by our support team. Since, there is an ongoing feedback loop between GenAI systems and our team agents that enable continuous learning and improvement.
Second, adoption remains challenging as teams naturally gravitate toward familiar workflows. Despite proven benefits, transitioning from conventional problem-solving approaches to GenAI-augmented methods requires dedicated change management efforts. We are running internal campaigns and training programs to equip our workforce with such advanced tools to enhance their productivity. We will continue to invest resources and time in building relevant skill sets and teams to drive PayU’s business success.
Looking ahead, how do you see the role of AI and machine learning evolving in the Indian fintech sector? What innovations can we expect from PayU in the AI space over the next few years?
We are bullish about the impact of ML and AI, especially GenAI on our business growth and, in fact, on the overall trajectory of the financial services industry in India. At PayU, while we continue to focus on building our core capabilities in 2025, including utilizing the latest LLM/GenAI technologies, to take full advantage of these advanced technologies, this year, we will also be doubling down on our GenAI initiatives and integrations to streamline operations, improve customer support, and drive efficiency.
For instance, one of our key initiatives is to further enhance merchant support through AI-powered email classification and response models, which currently address 30-40% of merchant inquiries instantly. Our goal is to automate 60-70% of tickets across emails, chatbots, and self-help tools, significantly improving response time and accuracy.
As mentioned, we are also particularly excited about exploring Large Language Models (LLMs) for code generation and review, aiming for 65% accuracy and a 50% reduction in Level 1 code review time. We also intend to leverage GenAI in the digital lifecycle management of merchants, in executing contextual cross-selling strategies, and in forecasting merchant revenue and churn.
Bhavya Bagga, APAC Media
Summarize with:
