A small business owner in Pune needed ₹4 lakh fast. His bank quoted 23%. Cars24 approved him in 60 seconds against his Maruti at 13%. 

That kind of outcome doesn't come from good intentions alone. It comes from a willingness to test and iterate until you reach that level of performance—the moment a customer realises this is genuinely different. What we call an “aha moment” in a product. 

And it comes from getting there faster than anyone else can deliver value.  

In lending, especially, speed without accuracy makes little sense. The two must move together. 

So, what made it possible?  

Some context on why this matters 

In early 2025, Cars24 launched Loans24, a fully digital credit platform offering used car loans from 10.99%, personal loans, and credit score building. It is open to anyone with a car, not just Cars24 customers. 

Loans disbursed grew 38% year-on-year to ₹1,637 crore globally, and the product continues to expand into new geographies, borrower profiles, and use cases as it moves toward becoming an end-to-end vehicle ownership platform. 

But behind that growth is a less visible bet: that Cars24 can build a lending product that gets sharper the more it’s used. Not just faster, but smarter. One that learns from every approval, every rejection, and every signal the portfolio throws off. 

That requires constantly asking uncomfortable questions. Does this bureau cutoff actually hold for a Tier II self-employed borrower? Is this fraud signal reliable in a state you've just entered? Does tightening an eligibility check improve the portfolio, or simply turn away customers you haven't learned to read yet? 

You don’t find the answers by thinking harder. You find them by testing. And for that to work at the pace Cars24 is moving, the infrastructure behind it has to be worthy of it. 

The execution roadblock 

Cars24's risk team knew their portfolio well. They could tell when a cutoff wasn't holding up for a borrower type they'd just started serving, when a fraud signal was misfiring in a state they'd recently entered, and when the data was pointing somewhere the current policy hadn't caught up yet. 

That's the job when you're lending in new markets. You're always testing, always adjusting. The policy that worked last quarter may not be the right one this quarter. 

The problem wasn't judgment—it was execution. Acting on any of this meant going through engineering. Every policy change, no matter how small, needed a ticket, a sprint slot, and a release cycle. A change a risk professional could work out in an afternoon took four days to go live. Experiments that should have happened weekly were happening monthly, if at all. 

Over time, the cost accumulated across every dimension: 

• Policy updates took four days from decision to deployment  

• Production issues took a full working day to diagnose and fix  

• Customers waited nearly 5 days for decisions meant to take minutes 

For a lender trying to get better with every market it enters, being a few days behind on policy isn’t just an operational inconvenience. It directly impacts approval rates and portfolio quality and weakens every lending decision being made. 
 
The intervention: Sentinel AI by FinBox 

Risk teams are professionally skeptical. A new tool doesn't get adopted because it looks good in a demo. It gets adopted when it earns trust. 

Cars24 brought in Sentinel AI by FinBox,  the cutting-edge Business Rules Engine, to eliminate the dependency between policy decisions and engineering releases. All decision logic—underwriting rules, KYC checks, eligibility criteria, fraud signals, and portfolio controls—moved into a single visual workspace that risk and product teams could own and manage directly. Sentinel AI is a decisioning OS that integrates risk, fraud, and borrower experience into one seamless platform.  

Three things changed: 

• Execution speed: Risk teams could build, update, and deploy policy changes themselves using visual rule builders. No code, no tickets, no waiting. 

• Safe experimentation: No change goes live blind. Teams can run simulations against historical data, shadow-test new rules alongside active ones without affecting live decisions, and use champion-challenger A/B setups to roll out changes to 10% of traffic before full deployment.

• Full visibility: Policy performance, approval trends, risk signals, and portfolio health—all in one place—so issue detection became proactive rather than reactive. 

The governance layer mattered just as much as the speed. Maker-checker workflows, role-based access, complete version history, and instant rollback meant the team could move fast without destabilising production. That's what earned adoption and made Sentinel AI the system of record for policy management across the organisation.

The numbers tell part of the story. But the 87.5% reduction in issue resolution time means something beyond the metric. It's the difference between a team permanently in firefighting mode and one that finally has the breathing room to be proactive. 

As Shobhit Singh, Director & Head of Risk Management at Cars24, put it: "Moving policy logic out of application code was a turning point for us. It gave our risk teams real ownership and removed the delays that came with every engineering release." 

What it made possible 

The TAT improvement is measurable, but the real win is the level of autonomy the risk and business teams finally have. 

At Cars24, they’re running experiments that used to be structurally impossible. They can test a tighter bureau cutoff on a specific segment or validate a new fraud signal on live traffic without needing a single engineering sprint. 

Most lending teams in India just can’t move this way. It’s not for lack of ideas—it’s just that every experiment usually means a Jira ticket and a two-week wait. When the friction is that high, you eventually stop trying things. Policy logic falls behind the portfolio, and underwriting starts to lag. 

Cars24 is building Loan Against Car to be the fastest, cheapest cash loan in the country. It’s not just an auto loan with a new look; it’s a personal loan built from the ground up. 60-second approvals, low-teens interest, and zero collateral friction. 

That business owner in Pune didn’t get a better rate by accident. He got it because the infrastructure finally caught up with the ambition.