MakeMyTrip is using an AI-first operating model to manage volatile, multi-modal demand at national scale while holding margins and cash discipline.
In Brief
- AI now controls large parts of discovery, booking and service, lifting conversion and decoupling service scale from headcount growth.
- Multi-modal supply and curated lodging networks give the business levers to redirect demand as fuel, capacity and geopolitics shift.
- Disciplined marketing spend and growing ancillary revenue turn volatile volumes into expanding operating margins and strong cash conversion.
The Strategic Break: AI At The Centre of Demand Management
MakeMyTrip has moved from using AI as a bolt-on optimisation tool to treating it as a foundational layer for how demand is captured, shaped and serviced across its network. That shift matters because the demand load has more than tripled in four years: gross bookings have risen from $3.2 billion in fiscal 2022 to $6.6 billion in 2023 and $10.4 billion in 2026, compounding at roughly 34 percent.
At the same time, demand has become more fragmented and more volatile. Booking frequency per user is rising, with three to six trips a year across leisure, family visits and extended weekends becoming the new norm for the connected earning class. Travel is no longer a single, long-planned annual event; it is a series of shorter, often late-booked decisions. External shocks have been frequent, from Middle East conflict to capacity cuts and fuel spikes. Management notes that travel demand stayed resilient in the unimpacted months, but international segments swung sharply with each disruption.
Against that backdrop, MakeMyTrip has chosen to redesign its operating logic around an AI-native front end and a diversified, multi-modal supply backbone. The goal is not only to automate existing workflows, but to create a system that can absorb fast-changing demand and route it through the most resilient and profitable capacity available at any given time.
How AI Reshapes The Planning and Service Stack
The most visible change is the deployment of the AI assistant Myra as an end-to-end conversational interface. Myra now supports planning, booking and payment across flights, hotels, buses, trains, cabs and full itineraries, including multilingual voice. Over the last quarter it handled more than 50,000 conversations a day, scaling to over 80,000 per day in recent days. About 45 percent of that volume comes from Tier 2 and smaller cities, 70 percent of queries are in a hybrid Hindi–English vernacular, and 10 percent of voice volume already flows through regional languages across seven added Indian tongues.
Operationally, this turns AI into the primary orchestration layer between demand and supply. Almost 15 percent of Myra conversations now occur at trip planning stage, not just at booking or support. Users who interact through Myra convert around 10 percent higher than those on traditional filter-led journeys, and in Q4 the assistant directly supported over 200,000 bookings. This changes the planning cadence: rather than react to search and booking logs alone, the platform can observe intent earlier and in richer detail, feeding forecast, pricing and assortment decisions.
The same pattern appears in customer support. Across flights and hotels, around 55 percent of call centre queries are now resolved by digital voice agents. On the bus business redBus, AI chatbots have delivered about 33 percent efficiency gains in support, and voice bots are being rolled out to replace legacy IVR. Regional language users show roughly twice the engagement of English-only users, with about 6 percent of total queries already via voice.
Inside engineering, 60 to 70 percent of new code is now written with AI tools. That does not change the supply network on its own, but it shortens the cycle from identified need to deployed capability, which is critical when demand and supply conditions move quarter by quarter.
In operational terms, this kind of AI-first stack typically requires:
- Clean, unified demand and supply data as the base for training models and powering conversational journeys.
- Clear allocation rules so that AI-led recommendations do not conflict with revenue, capacity or service thresholds.
- Tight integration between AI interfaces and booking, payment and support systems to avoid handoff failures.
MakeMyTrip discloses outcomes rather than architecture, but the scale of automation in code and call centres and the measured lift in conversion indicate that these foundations are in place at least for the core India business.
Multi-modal Supply as a Volatility Buffer
The AI layer sits over a deliberately broad supply network. On lodging, the platform now offers more than 100,000 accommodation options across over 2,050 cities, with room nights sold for 12,000 properties for the first time in the last year. Domestic hotels are growing faster than industry; in a quarter when external research indicated slightly negative occupancy year-on-year, MakeMyTrip recorded 15.2 percent volume growth in accommodation, with stand-alone domestic hotel volume up 15.5 percent. On a single January weekend the platform crossed 200,000 domestic room nights in one day.
At the same time, the company is promoting underpenetrated corridors such as the Northeast and spiritual destinations, and combining pilgrimage with leisure itineraries. Homestays are being integrated with additional signals such as whether quick commerce and food delivery are available and whether an on-site caretaker is present. That reduces perceived risk in more fragmented supply and encourages usage beyond standard urban hotels.
On transport, the network has become explicitly multi-modal. In the quarter, domestic flown passenger traffic declined 1.5 percent year-on-year and international fell 6 percent, with West Asia conflict and fuel and currency pressures weighing on capacity and fares. MakeMyTrip maintained leading domestic air market share at 30.8 percent and even gained 0.2 percentage points, but it did not rely on air alone to protect volumes.
Private bus inventory was expanded to an average of 46,000 daily schedules in the quarter, supported by a revamped route suggestion module that highlights corridors with unmet demand. Bus ticketing volumes grew 27.6 percent year-on-year in the quarter and 32.9 percent for the full year; intercity cabs grew over 20 percent. Bus ticketing adjusted margins rose 17.1 percent year-on-year to $41.1 million. Management links a 15.2 percent accommodation volume lift in the quarter directly to the availability of varied transport options, indicating that ground capacity and hotel utilisation are planned together rather than as separate verticals.
At network level, this is implemented through:
- Contracting and onboarding capacity across buses, cabs and hotels in corridors where air becomes expensive or constrained.
- Data-led guidance to bus operators on where to add or adjust routes.
- Cross-mode journey design that links bus and cab options into end-to-end trips in the front-end experience.
The result is a portfolio where demand can be steered from westbound international to domestic and eastbound travel, and from air to bus or cab, when shocks hit specific modes or regions. That is a structural hedge against airline capacity cuts, fuel volatility and regional instability.
Margin Discipline Through Mix and Cost Control
The AI-first and multi-modal choices sit inside a conservative margin posture. For fiscal 2026, IFRS revenue grew 10.7 percent year-on-year in constant currency, EBIT reached $156 million with 30.1 percent growth, and adjusted operating profit on gross bookings expanded from 1.71 percent to 1.82 percent. Management guides to 1.8 to 2 percent margins on gross bookings and has not raised that range despite recent expansion, citing the need for more stable travel conditions.
The levers behind that expansion are mix and cost discipline rather than simple volume growth. In the air business, adjusted margin rose 10.7 percent year-on-year in the quarter to $99.3 million, even though volumes declined because of disruptions. The gap was closed through higher ancillary attach and better unit economics per booking. In hotels and packages, adjusted margin grew 11.5 percent year-on-year in the quarter and 15.7 percent for the full year, skewing toward domestic hotels where control over supply and ancillaries is stronger.
Ancillary revenue is scaling quickly. Adjusted margin from other segments reached $25.4 million in the quarter, up 27.1 percent year-on-year, and $95 million for the year, up 37.1 percent. Marketing and sales promotion expense fell from 5.6 percent to 5.2 percent of gross bookings between the prior high season quarter and the latest quarter, yet adjusted operating margin held at 1.82 percent. That suggests more direct and repeat demand and more efficient targeting, supported by AI-driven personalisation and the conversational front end.
On cash, the business converted 97 percent of adjusted operating profit into operating cash flow in fiscal 2026, generating $182.5 million and ending the year with over $782 million in cash and equivalents. That level of conversion in a prepayment-heavy industry signals tight control over partner settlements and receivables.
The constraint is that these gains are being delivered at thin overall margins, and management is explicitly cautious. Targets remain anchored at sub-2 percent despite structural improvements, which implies that marketing spend, AI investment and supply expansion are being paced to avoid over-reaching in a volatile market.
Benchmark Contrast: AI as Operating Discipline, Not Just Labour Arbitrage
Across other experience-heavy networks, AI is often being used first to strip out labour hours. Leisure operators such as Lucky Strike report using AI systems over hundreds of locations to cut tens of thousands of frontline hours within weeks, while cruise operators such as Carnival emphasise AI and digital tools mainly as a way to keep unit costs growing slower than yields.
MakeMyTrip follows a more demand-centric pattern. The most visible metrics are not hours removed but conversations handled, languages supported and conversion uplifts achieved. Cost efficiencies in call centres and engineering are material, but they are framed as secondary outcomes. The primary role of AI is to capture, interpret and route demand in a way that keeps utilisation, ancillaries and margins resilient through repeated external shocks.
That contrast is important for any network operating with high fixed-cost infrastructure or contractual capacity. AI deployed primarily for workforce reduction yields a one-time cost reset. AI embedded as the main orchestration layer for demand and supply changes how planning and allocation decisions are made every day.
What This Operating Model Now Enables
The combination of an AI-native front end and a diversified, curated supply network gives MakeMyTrip a set of levers that are not tied to any single mode, corridor or customer segment. It can see intent earlier, in more languages and from more locations; it can redirect that intent across air, bus, cab and hotel capacity; and it can monetise each trip more fully through ancillaries, even when core volumes are disrupted.
The limiting factor is that this is being run within a tightly defined margin band, under continuous external pressure from fuel, currency and geopolitics. The operating model now enables faster, more granular responses to that volatility, but it does not remove it. The discipline sits in how AI, supply expansion and marketing are sequenced so that growth in gross bookings continues to translate into expanding, cash-backed margins rather than chasing volume for its own sake.
