MakeMyTrip's Myra 2.0 AI travel assistant: 85,000 daily conversations and over 50% of post-booking queries resolved, in the company's own words

The problem

MakeMyTrip Limited, described in its own filing as “India’s leading travel service provider”, sells flights, hotels and holidays to a market where a growing share of demand comes from smaller cities and from customers who are more comfortable in Indian languages than in English [source]. The operational question a buyer would ask of any travel platform is whether a conversational AI assistant can actually carry booking and support volume at scale, rather than sitting on the marketing site as a demo.

What was built

In its Q1 FY27 results MakeMyTrip said, “We launched Myra 2.0, the next version of our AI-powered travel assistant, which is now an end-to-end conversational booking platform” [source]. The company described the scope of the system in the same filing: “Customers can now search, compare, ask contextual questions, upload documents, and complete bookings, including agentic payments, through a single conversational interface by voice if preferred, across eight Indian languages” ([source]). Independent trade press summarised the same design, reporting that Myra 2.0 “runs an entire booking end-to-end inside a chat, including agentic payments” and quoting group CEO Rajesh Magow’s description of the interface [source].

The outcome

MakeMyTrip put two operating figures against Myra. On usage, the filing states, “Myra engaged in over 85,000 daily conversations in Q1 FY27, of which more than 45% originated from Tier II and Tier III cities in India” [source]. MediaNama corroborated the volume and geographic split, reporting that Myra “handled more than 8 million conversations during the quarter, including over 3 million in June alone, with more than 45% of usage coming from Tier 2 and smaller cities” ([source]). The same report frames the scale-up as steep, “from 25,000 conversations a day in the quarter ended September, as MediaNama reported, to more than 85,000 a day now” [source].

On support efficiency, the filing states, “Myra is also driving meaningful efficiency gains post-booking, resolving over 50% of post-booking queries in relation to flights and hotels during Q1 FY27” [source].

Two honesty caveats sit on the support figure. First, the scope is narrower in the primary source than in the secondary one, and the two are shown here rather than merged: the SEC filing scopes the 50% to “post-booking queries in relation to flights and hotels”, whereas MediaNama paraphrases Magow’s earnings-call remark more broadly, quoting him that “our AI-powered customer support bot independently resolves over 50% of customer calls” [source]. The narrower SEC wording is the one this page relies on. Second, and load-bearing for the whole page, the weakest supporting source is MediaNama, which is independent trade press but repeats MakeMyTrip’s earnings-call figures rather than measuring them, so it corroborates that the company said these numbers, not that they are independently true ([source]).

Why this caps below green

Every load-bearing figure on this page is MakeMyTrip’s own, stated in a regulated SEC filing and on an earnings call. A regulated filing is why these claims are documented rather than marketing, but a company reporting favourable operating metrics about its own assistant is not the same as an independent audit of that assistant. Under TIN’s rule, green never rests on the subject’s own numbers, and there is no non-MakeMyTrip measurement of Myra’s resolution rate or conversation volume in the public record. The honest ceiling is corroborated.

Related case files

• Octopus Energy’s ‘Magic Ink’ AI drafting support replies — the closest parallel: a company reporting its own AI’s share of support handling, the same first-party-attribution shape as Myra’s post-booking resolution figure.
• Windstar Cruises and Cresta’s chat agent containing 70% of chats — the same beat: a travel operator putting a first-party number against an AI assistant’s share of support conversations, useful for comparing self-reported AI containment rates in travel.
• Klarna’s AI assistant and its customer-service figures — the archetypal first-party AI-assistant customer-service metric, and a case that shows what more of the public record it took to move such a claim off a self-report.