NEW DELHI : Traveltech platform Ixigo (Le Travenues Technology Ltd) is betting on voice-enabled payments as part of its broader push towards a voice-first travel experience, although the feature is currently undergoing compliance review.
NEW DELHI : Traveltech platform Ixigo (Le Travenues Technology Ltd) is betting on voice-enabled payments as part of its broader push towards a voice-first travel experience, although the feature is currently undergoing compliance review.
Unlike click-based transactions, which automatically generate trace logs, voice interactions require a reliable way to verify that a customer explicitly authorized a transaction while also ensuring privacy.
Unlike click-based transactions, which automatically generate trace logs, voice interactions require a reliable way to verify that a customer explicitly authorized a transaction while also ensuring privacy.
"The technology itself isn't difficult to build. We want to do it in a way that doesn't lead to inadvertent payments. That's why the safeguards matter,” said Aloke Bajpai, co-founder and group chief executive officer (CEO), Ixigo.
Bajpai said the company wants to ensure every voice transaction is backed by verifiable records proving that the customer explicitly instructed the platform to complete the booking and payment. “The entire payments ecosystem will eventually develop standards for voice-based payments. Once those norms are in place, users could complete transactions without lifting a finger, simply by talking,” Bajpai added.
Ixigo has also launched its AI-native app and multimodal voice assistant, TARA. “If you don't have this kind of voice assistance at the top of the funnel, a customer would likely drop off and go to an AI platform or elsewhere to ask that question. But today, they can ask that question while making the decision right on our platform. As a result, we are able to reduce drop-offs, which means we are seeing some improvement in conversion,” said Bajpai.
Next frontier
The voice AI ecosystem is emerging as the next frontier, with funding increasing from $2.71 million in 2021 to $13.9 million in 2025, according to data platform Tracxn.
“Navigating apps remains a challenge for many people. Voice makes technology more accessible by letting users interact in a language they understand, unlocking digital services for anyone who can speak,” said Kumar Abhishek, founder and chief executive of Bengaluru-based fintech company ToneTag, which enables voice-first digital transactions.
There are multiple ways in which voice AI works. The simplest way is by giving a nudge. For example, if a person's loan payment is due, it might call and say, “Your loan is due, and I'm sending you a link,” before sending a payment link, said Ganesh Gopalan, founder of Gnani.ai, a voice AI company, which is also part of the IndiaAI mission.
“But voice AI has evolved beyond that. The more common approach now is an end-to-end, multimodal payment journey during a voice AI call. For instance, if a customer has missed an EMI payment, the AI agent informs them that the payment is due, sends a payment link through WhatsApp, SMS, or email, and stays on the call to guide them through the payment process if needed,” Gopalan added.
Strict guardrails
However, experts caution that agentic AI-based payments still require robust guardrails before they can be widely adopted in the country. This is where the consent of the customer becomes important.
“The customer must first give consent and authorize the capability. During a call, if they confirm they want to make the payment, the system captures that intent, verifies the amount, and processes the payment. We are careful to ensure the technology is not misused,” Gopalan said.
Currently, India does not have a specific framework for agentic AI payments. However, under the Reserve Bank of India's Digital Payments E-Mandate Framework, 2026, users must authenticate any payment mandate using an Additional Factor of Authentication (AFA), such as a one-time password or UPI PIN.
A similar approach is likely to apply to AI agents, add experts. Users would first have to authorize and authenticate an AI agent to make payments on their behalf. They would also need to set limits, such as how much the AI can spend and how long the authorization will remain valid.
“This is similar to the existing framework for recurring payments, as well as trigger-based buy orders, where the user provides consent upfront, and the system automatically executes transactions once the pre-defined conditions are met,” said Ravi Goyal, partner at law firm Scriboard.