From high-touch lender to fin-AI company
Speaking to ET Now, Agarwal said the company has deployed two Nvidia chips to run proprietary AI models internally, marking a shift from a traditional high-tech, high-touch lending model. The results are already visible in the numbers: cost of funds has dropped from around 12% three years ago to 10.3%, beating the company’s own 10.5% guidance, while yields have simultaneously risen roughly 10 basis points to 17.04%. Agarwal credited this combination — falling costs and rising yields — alongside ongoing AI-driven efficiencies for the company’s ability to defend its margins going forward.
Asset quality: A “collection first” philosophy
On concerns that rapid loan book expansion could hurt asset quality, Agarwal was firm that Paisalo treats itself as “a collection business first and a lending business second.” He pointed to a multi-decade track record since the company’s 1996 listing, during which asset quality stayed below 2% except during the Covid years, when it briefly rose before being brought back down within roughly six quarters. The company’s underwriting model functions more as a rejection filter than a disbursement engine, he said, with incentive structures across the organization weighted equally toward collections and growth.
No fresh equity needed for expansion
Despite pursuing aggressive growth, Agarwal said Paisalo does not need new equity capital, citing a comfortable 35% capital adequacy ratio and underleveraged 2.2x debt-to-equity position. Instead of diluting existing shareholders, promoters have been steadily raising their stake through open-market purchases — now at 46-47%, up from a low of about 26% four years ago. The company also has a $50 million foreign currency convertible bond outstanding, of which $44 million remains unconverted at a strike price of roughly Rs 48. With the stock now trading above that level, Agarwal expects meaningful conversion activity, and possibly full conversion, within the current financial year.
Four pillars driving 24-25% loan growth
Agarwal outlined four growth levers underpinning the company’s three-year AUM, revenue, and profit-doubling target: the shift to AI-led operations, aggressive distribution expansion, new product launches, and continued cost-of-capital optimization. Distribution has grown more than fivefold since 2017, expanding into 12 new states in the last three to four years to reach 5,299 distribution points across 22 states. Six new products were launched in the most recent quarter alone.
AI is already doing the heavy lifting
The scale of AI adoption at Paisalo is substantial. In just two quarters, the company processed around 160,000 loan applications through AI-enabled onboarding, alongside 125,000 servicing cases and 225,000 risk management cases handled by AI systems. It has also run roughly 250,000 quality checks across audit, credit and operations functions using AI, while scaling from zero to two AI bots and five outbound voice bots handling 350,000 multilingual calls daily in Hindi, English, and Marathi.
Opex to stay elevated near-term, then ease
Agarwal acknowledged that operating expense ratios will likely hold steady or edge up in the medium term as the company continues investing heavily in in-house AI and IT infrastructure, including a revamped sourcing app and business correspondent platform expected later this year. However, he expects opex efficiencies to materialize meaningfully over the longer term as AI-driven automation scales across the organization.
-1024x683.jpg "Florida Roads Become a Battleground for Illegal Immigration")
