Cars24 bets on OpenAI agents to reshape used-car buying

The News

Cars24, one of India's largest online used-car marketplaces, has gone public with the scale of its bet on generative AI, and the numbers are striking. In a case study published by OpenAI, the company revealed that its voice and chat agents now handle more than one million conversation minutes every month, spanning car buying, selling, financing, follow-ups and post-purchase support.

The headline operational win is lead recovery. Cars24 says the agents help it reclaim roughly 12% of leads that would otherwise have gone cold, re-engaging buyers who typically drop off during a multi-day decision window. The company has also built a consumer-facing assistant it calls CarGPT, which fields test-drive bookings and recommends cars to shoppers.

The rollout is not confined to the storefront. Around 600 employees across the central organisation now use ChatGPT Enterprise and OpenAI's Codex coding tool, with the company reporting daily active usage in the region of 85 to 90%. Cars24 operates in India, the UAE and Australia, and its leadership frames the purchase of a car in India as a journey rather than a single transaction that agents can accompany from discovery to delivery.

Why It Matters

This is a data point in a much larger story: enterprise AI in India is moving out of the pilot phase and into daily production. For years, Indian corporates ran cautious proofs of concept while budget holders waited for proof of return. A million billed conversation minutes a month, tied to a concrete recovery metric, is exactly the kind of hard number that unlocks the next round of investment.

The closest reference point is Klarna, which in early 2024 claimed its OpenAI-powered assistant was handling two-thirds of customer service chats and doing the work of 700 full-time agents. That disclosure reset expectations for what a single deployment could absorb. Cars24's announcement does something similar for India: it reframes agents as revenue infrastructure rather than a cost-centre chatbot bolted on at the end.

Indian Angle

Because Cars24 is itself an Indian company, the sharper question is what its playbook means for the market around it. Start with the financing angle. Cars24 runs a sizeable auto-lending operation, and pushing AI agents into the loan funnel puts it squarely inside the RBI's tightening expectations on digital lending, from clear disclosure of terms to grievance redressal. Any Indian lender automating conversations at this scale will have to prove that an agent's promises are auditable and that a human can be reached when a deal goes wrong.

Then there is competition. Rivals such as Spinny and CarDekho now face a public benchmark. If a 12% lead-recovery lift is genuinely achievable, it becomes a number boards will ask their own teams to match, accelerating AI spend across the sector.

Finally, cost and talent. Indian developers building on frontier models pay in dollars while earning in rupees, so token efficiency matters more here than in Silicon Valley. Cars24's willingness to put Codex in the hands of hundreds of engineers signals confidence that the productivity gain clears that exchange-rate hurdle, and it sharpens the domestic race for engineers who can ship agentic systems in production.

FAQ

What exactly did Cars24 deploy?

Cars24 built OpenAI-powered voice and chat agents that handle buying, selling, financing, follow-ups and support, alongside a consumer assistant called CarGPT for test-drive bookings and recommendations. Internally, around 600 staff use ChatGPT Enterprise and the Codex coding tool, with reported daily active usage of roughly 85 to 90%.

How big is the AI workload?

The agents process more than one million conversation minutes every month across Cars24's markets. The company also credits them with recovering about 12% of leads that would otherwise have lapsed during the multi-day window buyers take to decide, which is the metric most directly tied to revenue.

What does this mean for Indian regulators?

Automating financing conversations at scale intersects with the RBI's digital-lending framework. Lenders using agents must ensure loan terms are disclosed transparently, that grievance redressal remains accessible, and that automated interactions are logged and auditable, so a customer is never trapped in a loop with no human recourse.

Where can I read the original announcement?

The case study was published on OpenAI's website, and it sets out the deployment details, the operating markets and remarks from Cars24's leadership. The link to the full write-up is in the attribution note below.

This story was reported by OpenAI. Read the full original coverage at OpenAI.