MakeMyTrip (MMYT) AI Voice Bot Handles About 50% of Calls Unaided

MakeMyTrip (MMYT) told its 2026-08-03 earnings call that its AI voice bot now handles almost about 50% of all inbound customer calls with no human intervention, and for the first time linked that share to what it pays outsourced call-center agents [1]. On the 2023-10-31 call, post-sales service was still only an efficiency scenario management had identified [2].

MakeMyTrip Limited (MMYT) is an Indian online travel platform for flights and hotels, plus buses, trains and cabs; it owns the goibibo and redBus brands, and its size is measured in gross bookings and revenue. Travelers who call after booking reach agents in outsourced call centers, a service the company buys. The AI sits at the front of that phone queue: it answers those calls directly, understands the request and closes it out in the same call, while complex and rare problems still fall to an outsourced agent and a second layer inside the contact center assists human agents with information. Its scope is post-sales service for flights and hotels on the MakeMyTrip brand only; Myra, the traveler-facing conversational assistant, and the redBus service bot are separate, and their numbers cannot be borrowed across them.

How the application evolved

It took close to three years to go from an idea to a system that answers the phone. At the 2023-10-31 call there was no product: management listed post-sales customer service as a generative AI efficiency scenario, imagining information assist for call-center agents on complex cases, or a bot in the journey to filter generic queries, with no timetable and no cost or headcount target [2].

Deployment began on 2024-10-23, walking the post-sales journey scenario by scenario, with the CFO tying operating leverage on fixed costs to the generative AI rollout and cautioning up front that full P&L reflection was a quarter or two away [3]. On 2025-01-23 it first put post-sales cost growth and automation in one sentence, but that sentence is corrupted in the transcript and bundles AI with general automation, technology and customer-experience investment, and cost discipline, in a peak travel season quarter; it shows only that the cost story started early [4].

The executor arrived on 2025-10-28: a generative AI voice agent integrated with the telephony system went live and answered real customer calls directly for the first time, handling date changes, web check-in and cancellations. That period's result was still qualitative, with calls handled without a person increasing but no share and no cost figure [5]. 2026-01-21 produced the first quantified rate, AI voice and chatbots autonomously resolving about half of customer queries across flights and hotels [6]; on 2026-05-19 it gave 55%, on a different basis, call-center flights and hotels queries [7].

Where the application stands now

On 2026-08-03 the CEO put the voice bot at almost about 50% of total calls, frictionless, completely on its own, with no human intervention; prepared remarks in the same document say more than 50% of customer calls. The denominator is all calls, different again from the two before it, so the three printed shares do not form a rising curve [1].

The same call first tied the share to cost: the CEO said it is reducing outsourcing cost quite significantly, that this shows in the overall expense improvement, and that the business is growing while outsourcing cost has not grown at all [1]. Calls the AI takes consume no outsourced agent hours, so the company need not add agent capacity in proportion to bookings. The matching metric is the outsourcing fee sub-item under other operating expenses, and through it operating margin; the chain stops there.

The company reports under IFRS and its statements carry no SG&A line of the kind management refers to, disclosing cost of services, personnel expense, marketing expense, other operating expenses, and depreciation and amortization. The large line holding the outsourcing fee grew in the quarter and the full fiscal year, and its share of revenue rose against the prior fiscal year, because it also absorbs the AI's own web hosting and technology maintenance costs, an increase far larger than the small decline in outsourcing. Costs are mainly incurred in rupees and reported in US dollars, with no constant-currency basis on the expense side; rupee depreciation alone is enough to produce that sub-item's slight decline.

The CEO capped it in the same call: at this stage the incremental cost of AI and the benefit it brings almost offset each other [1]. The share measures how many calls the system caught; it supports a directional statement but cannot be turned into an amount, a basis-point figure or a margin gain.

What to watch next

MakeMyTrip is past the threshold where most contacts are handled without a person, so the question is no longer the take rate but whether the agent hours saved reach the expense statement. So far the answer is half and half: the outsourcing sub-item did move from double-digit growth to a slight decline, and the company listed it as a reduction item for the first time this quarter, but the operating expense line containing it is still growing. Customer-service AI is a realized operating fact for MMYT and an unrealized margin fact.

What would most change that reading is a quarter-by-quarter series for the outsourcing fee, disclosed on a constant-currency basis or directly in rupees; without it, that small decline cannot be separated from rupee depreciation.

Application assessment

• MakeMyTrip-brand post-sales customer-service automation - GenAI chatbots and telephony-integrated voice agents that resolve flights and hotels service contacts without a human agent, plus agent-assist in the contact centre. | Business position: core operations | Adoption stage: limited production | Scope: single business unit | Value type: cost reduction

Sources

[1] Drillr - MMYT earnings call - 2026-08-03

"our voice bot is handling almost about 50% of the total calls now, frictionless, completely on its own, no human intervention. That is the one which is sort of helping on reducing the outsourcing cost quite significantly. And that's sort of reflecting in our overall SG&A improvement, if you really see."

[2] Drillr - MMYT earnings call - 2023-10-31

"How could we, on one hand, either the sort of get some more productivity gains for the agents -- for the call center agents for the really sort of complex cases by providing a lot of the information or filtering a lot of the information using this technology, or we sort of look at gain introduction of bot at some level in the journey, which can filter the generic queries that come our way on the post-sales service queues."

[3] Drillr - MMYT earnings call - 2024-10-23

"But for it to sort of reflect completely on P&L, I think it will be a quarter or two down the line, we will see the full impact of that, and that is one area."

[4] Drillr - MMYT earnings call - 2025-01-23

"For instance, back bar investments in automation and artificial, our post sales cost for the quarter increased by just about 7% year-on-year compared to almost 26% year-on-year increase in bookings and revenue thus driving cost efficiencies."

[5] Drillr - MMYT earnings call - 2025-10-28

"We recently launched GenAI voice agent for our flights and hotels customers, which is designed to handle all customer queries received via calls and offer resolutions to the consumers in the same call."

[6] Drillr - MMYT earnings call - 2026-01-21

"Our AI voice and chatbots are now autonomously resolving about half of the customer queries across flights and hotels, significantly improving service scalability and efficiency in the system."

[7] Drillr - MMYT earnings call - 2026-05-19

"Similarly, AI is also driving meaningful efficiency gains on our customer service function. About 55% of our call center flight and hotels customer queries are being now resolved by digital voice agents."