New Delhi, July 21 -- Digital lending platform Fibe is deepening its use of artificial intelligence (AI) across credit underwriting, fraud detection, customer servicing and internal operations, as the company looks to scale lending while extending credit to customers with limited formal credit histories.
The company said its AI-powered risk engine can make lending decisions in under three milliseconds at the point of application, while machine learning models help detect fraud, automate Know Your Customer (KYC) processes and improve operational efficiency.
The fintech is also building what it describes as an institutional intelligence layer, codifying years of underwriting expertise into AI systems so that lending decisions are driven by consistent risk frameworks rather than individual judgement. "We did not set out to build an AI-first company. We set out to solve a genuinely difficult problem: how do you lend responsibly to millions of Indians who have limited credit history but demonstrably strong financial behaviour," Balakrishnan Narayanan, Chief Product & Analytics Officer at Fibe, told TechCircle.
"AI turned out to be the most effective mechanism to do this at scale and with the precision these borrowers deserve," he said.
The AI-led transformation comes as the TPG-backed fintech prepares for its next phase of growth. Fibe has reportedly filed draft papers for an initial public offering (IPO), with plans to raise capital to expand lending to India's growing middle-income consumer segment. The company is betting that AI-driven underwriting, automation and operational intelligence will help it scale while maintaining portfolio quality.
AI agents automate operations
Beyond customer-facing applications, Fibe has deployed AI agents across internal workflows, including data collation, process tracking, operational queries and preliminary analytics.
Narayanan said these AI-assisted processes have reduced turnaround times and manual coordination, allowing the company's operational capacity to grow alongside its loan book without a proportional increase in headcount.
"The more significant benefit has been scalability. As our loan book has grown, these AI-assisted workflows have allowed our operational capacity to scale without a proportionate increase in headcount or process complexity," he said.
He added that AI agents are designed to augment employees rather than replace them, with human oversight retained for credit approvals, exception handling and customer-sensitive decisions.
AI assistant monitors borrower behaviour
Fibe has also introduced an AI-powered conversational assistant called Fiora, which goes beyond answering customer queries by providing contextual assistance across the loan lifecycle.
Unlike conventional chatbots that primarily respond to FAQs, Fiora can proactively remind borrowers about repayments, explain loan charges in simple language and recommend products based on individual financial profiles.
Internally, the system analyses customer behaviour to identify early warning signals that may indicate future repayment stress. Those insights are then shared with collections and risk teams, while also feeding back into the company's credit models to improve underwriting quality. Narayanan said the AI assistant is always disclosed to customers as an AI system, aligning with transparency expectations under India's digital lending regulations.
Alternate data widens access to credit
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. The approach is intended to help lenders assess customers who may have little or no formal credit history but demonstrate consistent financial behaviour.
"The outcome, practically speaking, is that we can extend responsible credit to a segment that conventional underwriting would have declined, without loosening our risk standards," Narayanan said.
He added that all behavioural 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.
Governance to define AI-led lending
As AI becomes increasingly central to lending decisions, Narayanan believes governance rather than algorithms will emerge as the key differentiator among fintech companies.
He said Fibe subjects its credit models to explainability reviews and fairness audits, testing approval rates and repayment outcomes across customer segments, income groups and geographies before deploying them. Models that produce outcomes not supported by legitimate credit-risk factors are rejected, he said.
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 behavioural signals.
"The next phase of AI in Indian fintech will be defined less by technical capability and more by governance maturity-explainability standards, de-biasing methodology, consent architecture, and the ability to demonstrate to regulators that these systems behave as intended under adverse conditions," he said.
Published by HT Digital Content Services with permission from TechCircle.