Rajat Deshpande of FinBox discusses how AI-driven platforms are transforming credit risk assessment, speeding up loan processes, and broadening access for thin-file and new-to-credit borrowers in the evolving digital lending landscape.
Artificial intelligence is moving from a back-office experiment to a core part of digital lending, and Rajat Deshpande says the shift is already changing how lenders assess risk, process applications and serve borrowers. In an episode of the Analytics Insight Podcast hosted by Priya Dialani, the FinBox co-founder and chief executive described a market in which machine learning is increasingly used to process documents, analyse financial information and automate lending workflows, with the goal of speeding decisions without removing human control.
FinBox presents itself as an AI-native digital lending platform that gives banks, non-banking financial companies, fintechs and other businesses a modular, API-first system for launching and scaling credit products. On its website, the company says its tools include loan origination software, a decisioning engine called Sentinel AI and risk-intelligence features aimed at improving approval accuracy while protecting portfolio quality. The company’s broader pitch is that lending can become both faster and more flexible when data, decision rules and workflow tools are integrated into a single operating system.
A major part of that strategy is DeviceConnect, FinBox’s alternative-scoring platform, which relies on device-level intelligence and other non-traditional data sources to build what the company calls the FinBox Inclusion Score. According to FinBox, the system is designed to help lenders evaluate thin-file and new-to-credit customers by drawing on behavioural and financial signals such as banking activity, UPI and wallet usage, telecom records and utilities. That approach reflects a wider industry effort to broaden access to credit for borrowers who may not have enough bureau history to be assessed through conventional methods alone.
FinBox is also pushing further into automation with Atlas, its AI-driven loan origination system. The company says Atlas is designed to cut credit turnaround time from 21 days to one day by handling tasks such as document processing, validation and credit assessment with agentic AI, a term used for software that can carry out multi-step tasks with limited supervision. Deshpande also discussed governance and human oversight in the podcast, underscoring a central tension in the use of AI in lending: lenders want more speed and scale, but they also need clear controls, explainable decisions and safeguards around credit risk. The company says its technology is being built with input from lenders, suggesting it expects the next phase of embedded finance to depend as much on trust and compliance as on automation.
Disclaimer: This article is intended to inform and educate, not to recommend or endorse any financial product, investment or strategy. Please consider your own financial circumstances and seek professional advice where appropriate before making financial decisions.
