Chatbot TARA resolves 92% customer queries, AI agents handle 45% calls: Ixigo's Rajnish Kumar

Without AI, we would be inefficient and incur more cost for customer support: Ixigo’s Rajnish Kumar

The bots are clocking in—but are humans clocking out? In some areas it seems that is the case as one of the top travel technology platforms Ixigo, despite managing lakhs of customers, requires a fraction of manpower to manage daily support calls.

"Over the past year, our team grew from just 488 to 509, only 21 new hires—despite scaling operations significantly. We are handling more than 400,000 people every day across different modes of transport, and to handle their support, we don't even require a fraction of that number daily to manage support calls," said company's Director and Group Co-CEO, Rajnish Kumar.

He added that this has been possible because AI is deeply embedded in their culture and across functions, from marketing to HR (human resource). “From customer support to operations, AI helps us deliver efficiency at scale without proportional headcount growth.”

AI taking over human roles

He added, "Our AI (Artificial Intelligence) chatbot TARA resolved 2.11 million customer queries (in Q4 FY25)—work that would have required a significantly larger human support team, highlighting the scale and efficiency AI brings to our operations. If we didn’t have AI, we would probably be inefficient and incur a lot more costs on this function."

Today, nearly 92 percent of queries are handled by TARA and 45 percent flight-related calls are now handled by our AI voice agent, Kumar added.

He estimates 60-80 percent of the customer calls will be handled by AI by around next year. “It will get to a point where most of your customer interactions and customer support issues will be solved end-to-end by AI. Human beings would only be involved in very highly complex and emotional situations. When that happens, the number of people needed would be a fraction of what it is today.”

Along with chat, TARA is voice-based and Kumar highlighted the happening with AI at one end and the consumer at the other end is almost indistinguishable from human beings. “People talking to voice agents often have no idea that they're speaking to AI. It can converse in English, Hindi, and Hinglish as well, switching languages fluidly based on how the customer speaks. It handles interruptions and maintains low latency, ensuring a real-time conversation feel, which is a huge challenge for AI-based systems.”

Sharing some use cases, Kumar cited an example of their AI system calling hoteliers pre-emptively to reconfirm bookings for customers.

“In India, there’s a significant issue with customers getting bumped from hotel bookings. Although it's not above 10 percent, it’s in the high single-digit. Depending on the conversation, if the hotelier seems non-committal, these calls are then recorded and analysed by AI to assess the probability of bumping the customer, and then we take evasive actions.”

Enhancing human efficiency

Beyond pre-emptive booking confirmations, the AI system plays a broader role in enhancing human efficiency, Kumar said.

“There’s one part where AI handles certain things end-to-end, and no human being is involved, but even human beings are augmented with AI. We have internal AI systems augmenting human beings. Every call that is happening with human beings is automatically analysed by AI. That analysis is used for training customer support folks.”

Kumar said that training has become much easier with these technologies because it helps understand the kind of issues being faced during different calls. “You can do sentiment analysis and emotional analysis on calls, and based on that analysis, create feedback loops that help improve people. We also have systems where people can get real-time help and suggestions from AI. While they are talking to a customer, AI can suggest actions and things they should say to the customer.”

AI driving revenue

With a dedicated AI team divided into three parts —-efficiency, revenue, and disruption, Kumar is seeing AI driving revenue.

“At an aggregate level, we have a close to 30 percent attachment rate for our value added services. So, 30 out of 100 tickets sold have an ancillary in it, which drives revenue for us. All of these ancillaries, the way they are working and priced, are completely driven by AI.

The AI is optimising the P&L (profit and loss) for those products. We have seen a significant impact on how these products have created more uptake in terms of people using them and more revenue and margin for the business,” he said.

Shaping the future

For their value added services, Ixigo is making use of predictive technologies which Kumar thinks will shape the future of travel apps in India.

“One of the things we have used a lot in the past is data to create predictive algorithms to solve consumer problems. Predictive technologies are also used to create products and services that drive revenues for us. For example, ixigo Assured, which is the guarantee of a full refund. Similarly, Price Lock, the product that allows you to lock a fare by paying a small fee, is also based on predictive AI algorithms. If the price increases in the future, you only pay the locked price; if it falls, you pay the lower price.”

Travel apps of the future will look very different, they will be a lot less transactional and much more conversational and natural to use, Kumar added.

“Apps will become hyper personalised. One of the good things about AI is that it works much better with memory. Most models are trying to use past conversations to personalise responses, and travel is no different. Travel will also become extremely preemptive. Ideally, the app should know what the consumer wants even before they want it,” he said.