AI BriefingHow Cars24 is scaling customer conversations and building faster with OpenAI
Key point
Cars24 recovered 12% of lost leads while handling 1 million minutes of customer conversations per month with OpenAI agents.
Details
Cars24, India's largest AI-powered auto trading platform, built voice and chat agents using the OpenAI API. Given the nature of the Indian auto market, where it takes customers days to weeks to decide on buying or selling a vehicle, Cars24 tackled the challenge of delivering a consistent, high-quality experience across millions of interactions with AI.
Deploying agents across the customer journey
Cars24 built an AI agent that, when a buyer calls, asks about budget, family size, commute distance, and preferred vehicle type, then presents recommended models from the catalog and guides them through test drive booking and financing options. Before the test drive, it handles visit confirmation, suggests alternative vehicles, and collects additional information related to financing. After the test drive, it checks the customer's decision and proceeds with next steps, and after purchase, it supports feedback and after-sales service inquiries.
On the seller side as well, agents handle vehicle information collection, inspection scheduling, sending notifications, rescheduling appointments, and providing market information. In particular, AI agents re-engage leads that used to drop off after 10 days, recovering more than 12% of them.
Expanding AI across the entire development organization
Cars24 deployed ChatGPT Enterprise and Codex to about 600 employees in its central organization (engineering, finance, legal, marketing, operations), and daily active usage reaches 85-90%. Product managers write Linear tickets with Codex, and engineering teams tag Codex on bug reports to automate tasks. Codex summarizes GitHub activity and posts team updates, reducing the number of standups.
Cars24's finance and investor relations department uses Codex to extract and analyze data from systems and automatically generate investor reports, and has also built a workflow that reviews purchase requests and purchase orders above a certain amount to detect anomalies and auto-approve them. Some teams have built their own "chief of staff" agents connecting Slack, Gmail, and WhatsApp, automating communication and scheduling as well.
Results: a 50% increase in customer support resolution rate and an 80% reduction in processing time for key workflows.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.