Fibe is accelerating its digital lending platform with artificial intelligence, aiming to scale operations while extending credit to customers with limited formal credit histories.
Real-time decisions and internal automation
The company reports its AI-powered risk engine makes lending decisions in real-time at the point of application, while machine learning models detect fraud and automate Know Your Customer (KYC) processes. Beyond customer-facing tools, Fibe has deployed AI agents across internal workflows to handle data collation, process tracking, and operational queries. By codifying years of underwriting expertise into AI systems, the firm ensures lending decisions are driven by consistent risk frameworks rather than individual judgment. This institutional intelligence layer helps standardize risk assessment across the organization.
Read Also: Firms drop AI pilots lacking returns, CEO says
Fibe has filed draft papers for an initial public offering, with plans to raise capital to expand lending to India’s growing middle-income consumer segment. The company is betting that AI-driven underwriting and automation will help it scale while maintaining portfolio quality. The AI-assisted processes have reduced turnaround times and manual coordination, allowing operational capacity to grow alongside the loan book without a proportional increase in headcount. These agents are designed to augment employees rather than replace them, with human oversight retained for credit approvals, exception handling, and customer-sensitive decisions. Narayanan emphasized that the primary benefit is scalability, noting that operational capacity has expanded without a corresponding rise in process complexity.
Proactive customer assistance
Fibe has introduced an AI-powered conversational assistant called Fiora, which provides contextual assistance across the loan lifecycle. Unlike conventional chatbots that primarily respond to FAQs, Fiora proactively reminds borrowers about repayments, explains loan charges in simple language, and recommends products based on individual financial profiles. The system analyzes customer behavior to identify early warning signals that may indicate future repayment stress, sharing those insights with collections and risk teams. Narayanan noted that the assistant is always disclosed to customers as an AI system, adhering to transparency requirements.
Using alternative data to expand access
A key part of Fibe’s AI strategy is Persona AI, a framework that builds credit profiles using alternative data such as UPI transactions, utility bill payments, and income patterns alongside traditional bureau scores. This approach helps lenders assess customers who may have little or no formal credit history but demonstrate consistent financial behavior. Narayanan explained that the practical outcome is extending responsible credit to a segment that conventional underwriting would have declined, without loosening risk standards.
Read Also: Weekly roundup key stories from Aug 23-28
The company said it subjects its credit models to explainability reviews and fairness audits, testing approval rates and repayment outcomes across customer segments, income groups, and geographies before deployment. Models that produce outcomes not supported by legitimate credit-risk factors are rejected. All behavioral and financial data used for such assessments is collected with explicit customer consent and governed under the Digital Personal Data Protection Act and RBI’s digital lending guidelines.
Looking ahead, Narayanan expects AI to evolve from an operational efficiency tool into the core decision-making layer for digital lending, powered by India’s expanding digital public infrastructure, including the Account Aggregator ecosystem, UPI transaction data and behavioral signals. He believes governance rather than algorithms will emerge as the key differentiator among fintech companies, and that Fibe subjects its credit models to explainability reviews and fairness audits.
