He picked the part of finance nobody puts on a deck: the follow-up call, the missed payment, the after-hours note typed into a CRM by an agent on their fourth coffee. Then he pointed a billion of those interactions at an AI model and asked what it heard.
The Quiet Bet on the Back Office
Prodigal's office sits at 655 Castro Street in Mountain View, a few doors from the cafes where Y Combinator alumni run into each other and pretend they were just leaving. Inside, the company Shantanu Gangal co-founded with Sangram Raje in 2018 is doing something most fintechs avoid: building software for the people who chase money down, not the people who hand it out.
The thesis is plain. Lenders disburse loans in seconds. Then they spend years - and a lot of payroll - servicing them. Collections, payment reminders, dispute resolution, compliance auditing. The unglamorous middle of the lending lifecycle. Gangal's pitch on a Founders Unfiltered episode was characteristically blunt: "Lending is about collections, not disbursement."
Prodigal calls its core product the Prodigal Intelligence Engine - PIE for short, which is exactly the kind of dad-energy naming you'd expect from a founder who came up through IIT Bombay engineering and a Wharton MBA. PIE is trained on over a billion unique U.S. consumer finance interactions. Real calls. Real notes. Real disputes. Real apologies. The model learns the script and the subtext.
Layered on top: ProAgent, the AI agent that does the call work. ProNotes, which writes the post-call summary (now in Spanish). ProPay, the self-serve portal that nudges borrowers into a plan they can actually keep. Every product begins with Pro. Every product is aimed at the same metric: minutes saved per agent per shift.
On an ACA podcast, Gangal did the napkin math. Collections agents spend two to two-and-a-half hours of every shift writing notes, scheduling reminders, posting payments. At scale, that's roughly $8,000 to $10,000 per agent per year of pure wrap-up time. Prodigal's pitch is that AI takes that hour back. Humans take the harder conversation.
The Compounding Resume
IIT Bombay, then BCG & Blackstone
A B.Tech from one of India's most selective engineering schools led to early consulting and finance roles at Boston Consulting Group and Blackstone. Two firms that teach the same lesson from different ends: details cost money.
Wharton MBA
Then Philadelphia. An MBA from The Wharton School - the credential most engineers grab when they decide capital allocation is a fun problem too.
Fundbox, 2015-2018
Head of Data Science and Analytics at the small-business lending startup, where the operating reality of a credit book - models, risk, defaults, recoveries - became muscle memory.
From IITB to YC S18
Where the Hours Go
Gangal's most repeated stat on the podcast circuit - the wrap-up burden carried by every collections agent in America. It's the gap his product is built to close.
Source: Gangal's estimates on the ACA International podcast. 8-hour shift, illustrative.
The Shape of the Company
The Pitch, Reduced
Strip away the deck. Prodigal sells one thing: time. Take the dull, repeating, error-prone work that sits between a borrower and a payment, and give it to software trained on the largest pile of consumer-finance conversation data anyone has assembled.
The unsexy part is the moat. Compliance is brutal in this category - Reg F, FDCPA, state-by-state variations. Gangal's frequent appearance on the Receivables Roundtable series is not coincidence. He's spent years convincing risk officers that an AI agent can be auditable, that a call summary can be evidence, that a guardrail can hold.
On LinkedIn, he wrote in late 2025 about Prodigal's own hiring data: "From job snatcher to job creator." His argument: AI didn't shrink his teams; it changed what the teams did. More analysts, more solution engineers, fewer hours of typing.
The optimism is calibrated. It comes from a founder who has watched a hundred banks reject an AI pilot, then approve one, then expand it.
Three lines he keeps coming back to
"Lending is about collections, not disbursement."
"What makes our work exceptional isn't just the scale of our data; it's the domain expertise underpinning it."
"From job snatcher to job creator."
Things You Didn't Know
What He's Building Toward
Ask any consumer-finance executive what they'd buy if AI worked perfectly tomorrow, and most will name a version of what Prodigal is building: a single intelligence layer that listens to every call, writes every note, scores every account, drafts every dispute response, and knows the regulations cold.
Gangal's bet is that the company that wins this category will not be the one with the flashiest model. It will be the one with the most domain-specific data and the deepest relationships with compliance teams. He's spent six years stacking both quietly.
On a Receivables Roundtable episode he walked through Reg F implications of AI-driven outreach for an hour without raising his voice. That's the disposition. Soft-spoken, slightly amused, working a problem that isn't going to make a headline until the day it does.
Prodigal is past the cute phase. It's past the demo phase. Series A is closed and aged. The company is hiring against a thesis it's defended for years. And Gangal, the engineer-turned-consultant-turned-banker-turned-data-scientist-turned-founder, is finally just one thing on a business card: CEO.