Global Fintech Fest 2026 Reveals How AI Is Transforming Financial Services With Human Oversight

AI Integration Reaches Practical Stage in Financial Services

The Global Fintech Fest 2026, held in Mumbai from September 8-11, marked a turning point in how AI in financial services is being discussed and deployed1

1

. With participants from over 70 countries, the event centered on agentic AI, tokenization, and quantum computing, but conversations shifted from theoretical possibilities to concrete implementation challenges within the BFSI sector. Financial institutions, fintech companies, and regulators focused on where AI agents should operate and where human judgment remains essential.

Perfios Group CEO Nitin Chugh emphasized that complete automation is not on the horizon. While agentic AI will handle repetitive, low-level tasks like code generation, humans will always make final decisions1

1

. The role of professionals like underwriters is evolving rather than disappearing—where they previously processed 10 loan applications, AI enables them to handle significantly more. HyperVerge's business-loan underwriting suite demonstrates this division of labor, with agents pulling information from bank statements, GST filings, and income-tax returns, flagging gaps, and assembling credit notes in about a minute compared to two hours manually1

1

. However, lending decisions remain with credit managers, preserving human oversight in decision-making.

Visa Launches AI-Led Agentic Commerce Infrastructure

Visa positioned itself as the trust layer for India's digital payments ecosystem as it transitions toward AI-led agentic commerce2

2

. Suresh Sethi, Group Country Manager for India and South Asia at Visa, described India's digital payments journey as "a showcase for the world" and outlined how AI agents will eventually make decisions and execute transactions on behalf of individuals and enterprises. India's digital payment transactions increased from 6,365 crore in 2021 to 26,762 crore in 2025, while total value grew from INR 1,675 lakh crore to INR 3,144 lakh crore3

3

.

Visa is developing Visa Intelligent Commerce to enable trust-based interaction between websites and AI agents, including trusted agent identification, agent certification, and agent-led commerce journeys2

2

. The platform will help commerce sites become discoverable by AI agents while helping consumers identify more trusted agents. Visa is also building Visa Payment Pass Keys as a second-factor authentication mechanism. Sethi emphasized that trust cannot be bolted on afterward—it must be embedded in the architecture from the beginning.

Cyber Resilience Demands AI-Driven Defense Systems

Visa launched a whitepaper titled "Frontier AI: A New Era of Cyber Resilience" at Global Fintech Fest 2026, addressing how automated cyber threats require equally automated defense mechanisms3

3

. Subra Kumaraswamy, Chief Information Security Officer at Visa, revealed that exploiting a payment system weakness took about a year and a half in 2016 but now takes under an hour4

4

. India recorded INR 22,495 crore in cyber fraud losses across 2.81 million cases in 2025, up 24% year-over-year, with investment scams accounting for 76% of losses3

3

.

Kumaraswamy argued that security operations built around human approval for each step cannot keep pace with automated cyber threats. He proposed shifting from "human in the loop" to "AI in the loop" for security operations centers4

4

. The whitepaper introduces Mean Time to Adapt (MTTA) as a new metric for cybersecurity effectiveness, measuring the time between vulnerability discovery and remediation3

3

. Visa open-sourced the Visa Vulnerability Agentic Harness (VVAH) in June 2026 and expanded it on August 27, allowing organizations to use various AI models to hunt for vulnerabilities and generate patches4

4

.

Specialized AI Models Gain Traction Over Large Language Models

Financial companies operating under strict regulatory requirements are gravitating toward smaller, specialized models rather than deploying the largest available large language models1

1

. This approach provides greater control over sensitive data while managing costs more effectively. KFintech, India's largest Registrar and Transfer Agent, used open-weight models from Alibaba's Qwen to build specialized AI systems for handling large volumes of sensitive data. Mastercard's AI Garage built a large tabular model for transaction data, allowing the company to work with characteristics specific to its payments data. Revolut India developed PRAGMA, a family of transformer-based models tailored to its operations.

Companies like KFintech, Zeta, and Perfios are adopting agentic workflows only where absolutely mandatory to minimize AI spending1

1

. The industry recognizes that compute power carries significant costs, making it essential to understand which tasks need automation and which can be done manually. This cost-conscious approach reflects a maturing understanding of AI deployment in financial services.

Job Transformation Rather Than Elimination

Contrary to widespread concerns about job losses, Global Fintech Fest 2026 revealed an optimistic outlook for employment in the AI era. Revolut India CEO Paroma Chatterjee stated that making people redundant due to AI is a no-win situation, announcing plans to hire approximately 1,500 people in India in 20261

1

. She emphasized creating AI-targeted jobs and building workflows with AI embedded in the design thinking. Roles built around repetitive tasks will see the biggest disruption, while senior positions will increasingly handle higher-value work rather than routine processing.

Zeta's APAC CEO Ramki Gaddipati explained that the firm's coding teams have shifted from basic code-writing to reviewing AI-generated code, with AI agents handling the first level of coding1

1

. This transformation illustrates how job roles are evolving rather than disappearing, with continued emphasis on hiring and upskilling employees to work alongside AI systems effectively.