3
The digital payment ecosystem in India is developing rapidly. Recently, at the India AI Impact Summit 2026, Mastercard successfully demonstrated the first-ever ‘Agent Commerce’ transaction in India. Under this initiative, AI agents will be able to make payments on behalf of users in a secure and tokenized fashion. Mastercard aims to partner with AI companies, fintechs, and merchants to accelerate AI-based commerce in the Asia-Pacific region.
The first transactions were through Mastercard cards, which were issued by Axis Bank and RBL Bank. These tokenized agent purchases were made with the help of payment aggregators like Cashfree Payments, Juspay, PayU, and Razorpay. Participating merchants included major names like Swiggy, Instamart, Vi and Tira.
The system is based on Mastercard’s ‘Agent Pay Framework’, which defines standards for tokenization and security. In the future, users will also have the ability to choose the AI agent of their choice, allowing interoperability and large-scale usage.
Nitendra Rajput, senior vice president of Mastercard AI Garage, said that there may be some risks in allowing AI agents for financial transactions. The biggest danger is “misaligned optimization,” in which AI may prioritize speed and efficiency while compromising customer interests. The second danger is “cascading errors,” in which small mistakes due to automation can turn into big losses.
The third risk is that of over-reliance, where users become unduly reliant on AI decisions. To address these risks, clarity in systems, continuous monitoring, independent auditing, and human oversight will be critical.
In the present time, fraudsters are also using AI. Mastercard says their fraud models work on a “Mixture of Experts” approach, with multiple AI engines working together. These models are constantly evolving, keeping them one step ahead of fraudsters. Through this strategy, success has been achieved in preventing fraud worth billions of dollars.
With India’s digital diversity and diverse tools, local AI models are being created so that the system is not only responsive but also structurally robust.
Debates to data localization are reshaping AI systems in financial services. Mastercard says that privacy and regulatory compliance are non-compromising for them. This means that national data limitations will be respected in the training, monitoring and implementation of AI models.
However, when it comes to fraud, some patterns first appear globally, leading the company to adopt a hybrid model by combining global algorithms with local data.
Credit card spending in India reached ₹2.12 lakh crore in January, according to RBI data. Digital payments are growing rapidly, especially in small towns and villages. India has proved that digital payments can be made secure and reliable even on a large scale. This is why India has become a major testing ground for AI-based payment systems.
Share your reaction or leave a quick response — we’d love to hear what you think!
