Air India (AI) is deploying more than 30 artificial intelligence initiatives across its operations in a bid to cut costs, improve operational precision, and redefine the passenger experience following its privatisation under the Tata Group in January 2022. According to a PTI report cited by multiple Indian media outlets such as ETV Bharat, the airline’s Chief Digital and Technology Officer, Dr. Satya Ramaswamy, confirmed in April 2026 that these AI use cases hold the potential to deliver annual savings in the region of ₹100 crore, with measurable efficiencies already documented across contact centre operations, engineering, employee support systems, and pilot scheduling compliance as reported by Economic Times.
The disclosure came as part of Air India’s broader communication around its ongoing multi-year transformation programme, which the Tata Group has anchored in technology as the primary vehicle for recovery after decades of government-era underinvestment. Ramaswamy told reporters that privatisation effectively provided the airline with a “blank slate” for technological innovation, an opportunity it seized by designing its AI architecture from the ground up rather than layering new tools onto legacy systems.
Inside Air India’s ₹100 Crore AI Savings Target
The ₹100 crore (approximately US $12 million) annual savings figure that Dr. Ramaswamy cited is a conservative, composite projection across departments. The airline has constructed over 30 in-house AI tools addressing everything from aircraft maintenance prediction to on-time performance improvement, with each tool yielding incremental savings that aggregate into the headline target.
Ramaswamy made clear that cost reduction is the explicit mandate from leadership: he told PTI that the airline worked extensively with all its senior executives to identify AI programmes of the highest priority to each department, and that many of those priorities centred specifically on departmental cost reduction.
The savings projection sits within a broader financial context that underscores why efficiency gains matter acutely for Air India right now. According to Singapore Airlines’ (SIA) annual financial statements, Air India recorded a loss of approximately SGD 3.56 billion (roughly ₹26,765 crore, or US $2.8 billion) for FY2025/26 — the largest annual loss since the Tata Group’s takeover.
The airline attributes those losses to the compounding effect of unprecedented external shocks such as:
- Pakistan’s airspace closure in May 2025
- record fuel price spikes due to the conflict in Iran
- A forced reduction in international capacity
Against that backdrop, the AI-driven savings programme represents not merely an efficiency exercise but a financial imperative.
How AI.g Became The First Generative AI Customer Agent At Scale In Global Aviation
The most visible and most extensively documented element of Air India’s AI deployment is AI.g — the generative AI-powered virtual agent that Microsoft confirmed made Air India the first airline worldwide to deploy generative AI for customer service at scale. Built on Microsoft’s Azure OpenAI service, AI.g moved from internal development to live customer interaction within six months of Microsoft launching Azure OpenAI in November 2022. The speed of deployment was made possible precisely because the airline had no pre-existing chatbot architecture to dismantle or migrate.
Today, AI.g handles approximately 40,000 customer queries daily across more than 1,300 distinct topic areas — from seat changes and refund requests to baggage allowances and frequent flyer redemptions. Since its launch, it has resolved more than 13 million conversations with a 97% success rate, meaning only 3% of queries require escalation to a human agent.
AI chatbots have reduced inbound call volumes at Air India’s contact centre by 50%, directly reducing staffing demand for routine query resolution while enabling human agents to concentrate on complex, judgment-intensive cases.
Ramaswamy described the operational impact on staff in a Microsoft case study, noting:
“Employees are doing things that are more value-added. I think it helps with morale…because now they get to contribute at a higher level.” He also quantified the customer experience outcome bluntly: “AI.g now handles 97% of 4 million-plus customer queries. We’ve saved millions, but the real win is customers aren’t waiting — they’re served.”
Use of Predictive Maintenance and Engineering AI in Air India
Beyond the customer-facing layer, Air India has deployed artificial intelligence in engineering and maintenance as well. According to the Air India newsroom, the airline’s engineering division “benefits from predictive maintenance and improved spare-parts planning, while upgraded operational control centres now provide real-time network visibility and faster decision-making during disruptions“.
The airline maintains a central Customer Data Platform housing over 80 million customer profiles, enabling data-driven insights to flow across all operational touchpoints — including in-flight, via crew iPads. Predictive analytics have strengthened planning across maintenance, network performance, and scheduling, representing what the airline describes as a structural shift from reactive management to anticipatory execution.
These capabilities directly reduce unscheduled downtime by identifying component degradation patterns before failures manifest.
Ramaswamy also pointed to the deployment of AI tools for cabin supervisors and operations teams. Pilots use a tool called AI FlightPro as a digital operational companion, while cabin supervisors access CE Plus to manage onboard service in real time. Operations teams work through a unified platform called AI Smart, which provides a consolidated operational view across the network.
.jpg){kind=link}
How Generative AI Validated Air India’s Pilot Safety Compliance
One of the most consequential and technically nuanced applications of AI at Air India involved the implementation of revised Flight Duty Time Limitations (FDTL) norms for pilots — a regulatory requirement issued by the Directorate General of Civil Aviation (DGCA).
The historical process for implementing DGCA rule changes was manual and susceptible to error: veteran pilots would codify regulatory requirements into internal operational specifications, which would then be manually translated into software. This multi-step human chain created meaningful risk of specification gaps or incorrect software implementation.
Air India used generative AI to validate the three-way mapping between:
- DGCA rules
- the airline’s internal specifications
- the final software implementation —
This is a level of cross-referential validation that was simply not achievable before large language models became available. Ramaswamy stated that this ensures “the correctness of the implementation and its completeness. This was never possible before, but generative AI gave us the ability to do that.“
The airline also used generative AI to generate an exhaustive set of edge and corner test cases — scenarios where the software might produce an inadvertent regulatory violation under unusual operating conditions. Ramaswamy told PTI:
“In particular, we generated an exhaustive set of edge and corner cases to test the implementation, as they have the potential to result in violations which we want to very much prevent.”
The DGCA had previously imposed an ₹80 lakh financial penalty on Air India in March 2024 for FDTL violations discovered during a spot audit, making the AI-based compliance validation process directly material to regulatory risk management.
Agentic Frontier of AI That Air India Is Now Entering
Air India’s AI architecture operates across three distinct and increasingly autonomous tiers: predictive AI, generative AI, and agentic AI. Ramaswamy has been explicit about the strategic sequence:
- predictive AI handles pattern recognition and forecasting in maintenance and scheduling
- generative AI enables conversational interfaces and document analysis
- agentic AI that allows software systems to orchestrate complex multi-step workflows autonomously, with minimal human intervention.
Ramaswamy told Financial Express that the airline is now exploring “agentic coding” to build software systems internally at pace, describing agentic AI as “a very transformative shift as it allows machines to integrate with humans, understand context and act autonomously in ways that were not possible earlier.”
The goal, he stated, is to become “the world’s most technologically advanced airline.” One near-term application under trial is an agentic AI system for refund processing that could compress timelines from weeks to hours by orchestrating workflows across multiple back-end systems without requiring manual handoffs between departments.
The airline has also integrated AI.g into WhatsApp, enabling passengers to access real-time flight updates, boarding passes, seat selection, and baggage information through a direct chat interface at any hour. AI.g communicates in four languages — Hindi, English, French, and German — and learns continuously from unanswered queries, expanding its knowledge base over time without requiring manual reprogramming.
Air India’s AI Transformation in Context of Its Broader Financial and Fleet Recovery
The AI cost-saving programme does not exist in isolation — it is one element of a capital-intensive, multi-year transformation that the Tata Group has funded since completing the airline’s privatisation for ₹2,700 crore in January 2022.
Air India’s standalone revenue rose 13.5% in FY2024/25, and its standalone net loss narrowed 21% year-on-year, signalling improving operational leverage even as transformation expenditure remained heavy. The group now operates a combined fleet of close to 300 aircraft after absorbing Vistara in November 2024 and folding AirAsia India into Air India Express.
In FY2025/26, however, the airline faced a confluence of external shocks. Pakistan’s airspace closure following border tensions in May 2025 forced significant route detours on European and US-bound services. Record-high fuel costs and prolonged airspace restrictions prompted Air India to temporarily reduce nearly 100 international flights and suspend seven overseas routes, including the Delhi–Chicago service, cutting international capacity by as much as 27%.
Despite these external pressures, Singapore Airlines, which holds a 25.1% stake in Air India following the Vistara merger, confirmed that the airline continues to make progress in fleet renewal, retrofit programmes, customer experience improvements, and operational enhancements.
A comparison with a parallel development at the carrier is instructive. Air India also announced in mid-2025 plans to bring aircraft maintenance in-house as the carrier looks foward to moving away from its historical reliance on Air India Engineering Services Ltd. (AIESL) with support from Singapore Airlines.
That decision, accelerated in the aftermath of the fatal Air India Flight 171 crash in Ahmedabad in June 2025, directly reinforces the AI-enabled predictive maintenance programme: bringing engineering control in-house and using machine learning to anticipate failures before they occur are complementary pillars of the same operational safety strategy.
What Ramaswamy’s Vision Means for Aviation’s AI Trajectory in India
The scope and pace of Air India’s AI programme carry implications that extend well beyond a single airline’s cost structure. Microsoft’s Judson Althoff publicly described Air India as the first airline to scale generative AI for customer service globally. In June 2025, Adobe’s AI and Digital Trends India snapshot cited Air India as a leader in AI-driven customer transformation, with Ramaswamy stating: “AI is no longer a futuristic concept at Air India — it’s the engine powering our customer-first transformation.“
The airline’s approach — building a majority of its AI tools in-house rather than procuring off-the-shelf solutions — is a deliberate strategic choice. Ramaswamy said that agentic coding allows the airline to “pinpoint business requirements with high precision” and support departments cost-effectively, and that this model will become more consequential as AI systems grow more autonomous.
On the question of workforce impact, Ramaswamy was measured but direct: while he declined to predict reduced hiring, he stated unequivocally that “the nature of practically every role we have will change due to AI,” and that every employee would be empowered by AI tools. The airline’s 96% active daily usage rate across its seven employee-facing digital channels suggests that adoption has moved well beyond pilot testing into embedded operational practice.
The history of Air India as a carrier that shaped Indian aviation since its founding as Tata Airlines in 1932 is well documented. Our historical profile of the airline captures the arc from JRD Tata’s first Karachi-to-Bombay airmail flight to the nationalisation era and beyond, providing the institutional context against which the current technological re-founding must be understood.