Cars24, a used-car marketplace centered on India, has built OpenAI-powered voice and chat agents into its customer operations and shared the results[1]. In a used-car business where phone calls, document checks, and follow-ups can drag on for days, how far can AI carry the conversation and win back leads that had slipped away? Using the figures the company disclosed, here is a look at where the agents fit across the buying and selling journey.

What Kind of Company Is Cars24

Cars24 is one of the world's largest AI-native automotive marketplaces for buying and selling used cars, operating primarily in India. It also runs operations in the UAE and Australia, supporting the full car-ownership journey from discovery and financing to resale and post-purchase services[1]. By making pre-owned car circulation more efficient, it also helps extend the useful life of vehicles.

Much of India's used-car trade remains manual, heavily regulated, and fragmented across steps. Unlike typical e-commerce, buying or selling a car rarely finishes in a single session. Most of the process happens outside the app, and the calls, document checks, and follow-ups that come afterward can take days or even weeks[1]. The partnership between the two companies was announced in March 2026, and Cars24 has been searching for a way to handle this complexity without endlessly expanding its operations teams[2].

Voice and Chat AI Across Every Stage of Buying and Selling

Using OpenAI's APIs, Cars24 built voice and chat agents that cover buying, selling, financing, follow-up, and support. It started with the hardest part of the funnel — the middle and bottom, where conversation drives conversion[1].

When a buyer calls Cars24, an AI agent asks about their budget, family size, commute needs, and preferred car type, recommends cars from the catalog, books a test drive, and helps the customer explore financing. Before the test drive, it confirms the visit, suggests alternative cars if preferences have changed, and collects extra details for financing; after the visit, it checks whether the customer wants to move forward or look at another option. Even after purchase, the agents keep handling feedback, warranty questions, returns, and after-sales service[1].

The seller side follows a similar path. An AI agent gathers vehicle details, schedules an inspection, sends reminders, helps reschedule missed appointments, and gathers competitive insight when a car has already been sold elsewhere. For leads that used to drop out after around 10 days, the AI re-engages the customer, qualifies renewed intent, and returns them to the funnel when Cars24 can meet the price they are looking for[1].

The Results Show Up in the Numbers

According to Cars24, its AI agents now handle more than 1 million conversation minutes per month. Customer support resolution rates rose by 50 percent, and turnaround time across key service workflows fell by 80 percent. Of the seller leads that had previously been lost, 12 percent were recovered through AI-driven re-engagement[1].

A company representative put it this way: "Buying a car in India is a journey, not a transaction. For years, the experience depended on who picked up the phone. AI changes that. Today, we handle over a million conversation minutes a month through AI, giving every customer a high-quality experience at any scale."[1]

Codex Spreads Beyond Software Development

Cars24 has deployed the AI coding agent Codex across its entire software development lifecycle, treating it not as a standalone coding tool but as a participant in day-to-day work[1].

Product managers use Codex to create and refine tickets in the project-management tool Linear, and engineers tag Codex into bug reports to hand off defined tasks. Codex summarizes work across GitHub and shares it with teams, cutting down on status meetings. Its uses go well beyond engineering. Finance and investor-relations teams use Codex to pull numbers from internal systems, run analysis, and prepare reports. Another workflow reviews purchase requests and orders above a set threshold and auto-approves them when no anomalies are found[1].

Internally, teams have begun building their own "chief of staff" style agents that connect Slack, Gmail, WhatsApp, and other tools to manage communication, scheduling, and hiring flows. Cars24 has rolled out ChatGPT Enterprise and Codex to about 600 employees across its central organization, with daily active usage reaching 85 to 90 percent. Beyond engineering, teams in finance, legal, marketing, and operations are assembling their own workflows with AI[1].

Summary

Cars24 embedded OpenAI's voice and chat agents into India's used-car trade — where calls and document checks stretch on for days — and reported results including more than 1 million monthly conversation minutes automated, a 50 percent lift in support resolution, and 12 percent of lost leads recovered. Codex has also spread into finance and investor relations beyond engineering, and an "AI-native" way of working, where employees build their own AI tooling, is taking root inside the company. It is a case that suggests the key to moving AI from pilots to production is tying it concretely to specific business workflows.

Source: https://openai.com/index/cars24

Source: https://knowstartup.com/news/cars24-partners-openai-ai-automotive/