Cars24 Rolls Out OpenAI Voice and Chat Agents: 1 Million Minutes of Conversations a Month and 12% of Lost Leads Recovered

Indian car marketplace Cars24 has moved customer communication to voice and chat agents built on OpenAI models. The bots handle more than 1 million minutes of conversations a month and help recover 12% of lost leads. OpenAI published the case study on its blog in July 2026.

How Cars24 uses OpenAI agents

Voice and chat agents built on OpenAI models now handle the main flow of customer communications at Cars24: more than 1 million minutes of conversations every month. That volume, over 16 thousand hours of dialogue a month, is comparable to the full workload of a call center with roughly a hundred operators. Cars24 is a platform for buying and selling used cars, founded in India in 2015. A transaction here is expensive and takes a long time, so the speed and quality of responses directly affect revenue. The agents answer buyers' questions, help sellers with vehicle valuation, and bring back into the funnel customers who did not finish their applications.

Key figures from the case study, according to the OpenAI blog:

• Voice and chat agents handle more than 1 million minutes of conversations a month
• Cars24 recovers 12% of lost leads through automated communications
• Agentic workflows are being rolled out in teams across the company, not just in support
• Cars24 has operated in the used-car market since 2015

Why recovering 12% of leads matters

For a car marketplace, recovering 12% of lost leads means a direct revenue gain without a larger advertising budget. A lost lead is a customer who submitted an application but did not answer the phone or abandoned the process midway; in a traditional call center, these applications often never get followed up. Unlike a human operator, a voice agent can call thousands of customers at once, at a time that suits them, with no queue on the line.

The economics of voice agents in the auto business are more visible than for most chatbots: a car deal is worth thousands of dollars, so even a few percent of recovered applications pays for the deployment. In the case study, Cars24 ties the agents specifically to recovering lost demand, not just to cutting support costs.

What the agentic approach gives the company

At Cars24, agentic workflows reach beyond customer support: according to the OpenAI blog, the company is rolling them out in teams across the organization. The company calls these agentic workflows. An agent in such a process does not just answer a question; it carries out a chain of actions: it finds data, triggers the next step, and hands the result to a human.

"Cars24 uses voice and chat agents powered by

OpenAI to handle more than a million minutes of conversations a month, win back 12% of lost leads, and roll out agentic workflows in teams across the company," the case study on the OpenAI blog says.

The phrase "builds faster" in the case study's title refers to more than the customer-facing service: Cars24's internal teams use OpenAI tools to assemble workflows and internal products more quickly. This is a typical second stage of enterprise adoption: first the front line is automated, then agents move into internal operations.

What this means

Voice AI agents are no longer an experiment: at Cars24, they run at a scale of a million minutes a month in a business built around an expensive purchase, a car. For companies with a large flow of applications, the takeaway is simple: measurable returns come not from a "chatbot on the website" but from an agent embedded in the sales funnel that wins back lost customers. Case studies at this scale also show where OpenAI is steering the enterprise segment: from one-off pilots to agents in daily operations.