Inside the airport technology race

Editorial context

Indian airports are buying dashboards and digital twins. Sunil Ankalgi says the real transformation occurs when airlines, handlers, security and airport operators make decisions using the same timely information.

The customer is not always the buyer

A modern airport is an unusual technology customer. The airport operator pays for systems, airlines depend on them, passengers experience the consequences and government agencies control several essential processes. Sunil Ankalgi, who has worked with SITA and Amadeus and now leads WAISL’s business, begins with the uncomfortable truth that success looks different to every participant. An airline optimises turnarounds and schedule adherence; an airport authority is concerned with long-term infrastructure utilisation; security and border agencies must meet obligations not reducible to the fastest possible queue. A supplier that sells software to one stakeholder while neglecting the others risks building a technically competent platform that cannot change how the airport operates. Governance, access rules and operational authority matter as much as the code.

India cannot import the transfer-hub template untouched

Indian airport growth has been powered substantially by domestic origin-and-destination demand. Ankalgi contrasts this with airports in Singapore, Seoul and elsewhere where international transfers carry a different weight. That distinction is consequential for IT architecture. Transfer-heavy hubs depend on through baggage, immigration workflows, complex connection protection and coordinated airline banks. Indian airports upgrading terminals for rising local passenger volumes cannot assume that the same investments automatically make them strong international connectors. Digital systems should reflect the airport’s mix of domestic travellers, interterminal movement and multimodal connections. The debate about the physical Delhi air train is a reminder that some failures of integration cannot be solved by software. A perfect data feed still cannot move a passenger through a missing transport link.

An airport operations centre must make decisions

A room containing large screens is not necessarily an Airport Operations Control Centre. Ankalgi explains that a genuine integration goes beyond displaying siloed information from flights, bags, stands and passenger processing. It requires relevant actors to interpret the same status, anticipate knock-on effects and coordinate a response. He cites WAISL’s digital-twin-linked work at Hyderabad. The analytical distinction is between visibility, prediction and action. A dashboard can show that queues grew; a forecasting system can predict their growth; a functioning operations framework enables an airport to move staffing, revise stands or change flow in time to prevent a downstream disruption. Without decision rights and measurable intervention, ‘digital twin’ risks becoming a fashionable rebrand for business intelligence.

Turnaround minutes are not free capacity

One much-circulated industry claim comes up in the interview: saving ten minutes of turnaround time creates the possibility of several extra aircraft rotations. Ankalgi himself expresses doubt about the precision of the statistic. That caveat is essential. The gain depends on the operating day, aircraft and airline scheduling, slot restrictions, air traffic control, gate demand and whether the saved time is recurrent and bankable. Shaving ten minutes from a stand process cannot create a new runway movement at a slot-constrained airport. Nor do punctual turnarounds always create new flights: sometimes they create a better recovery buffer. The correct measurement is change in throughput or schedule resilience at the actual bottleneck, not multiplication of minutes by an arbitrary headline coefficient.

AI is in the terminal—but not running everything

Ankalgi distinguishes deployed applications of computer vision from more speculative automation. Queue management, baggage-flow monitoring and turnaround milestone prediction can exploit camera feeds and sensor data, while fully dynamic gate allocation across every constraint remains a much harder problem. The distinction reflects airport operations more broadly: detection is not dispatch. Predicting a delay is one problem; reconciling the interests of a hub carrier, ground handler and airport stand manager is another. Models require trustworthy time-series data and continuing supervision, because a wrong prediction can cascade across the operation. The case for AI should therefore be made through disruption minutes saved, missed connections avoided and asset productivity gained, rather than the number of algorithms deployed.

Biometrics remain fragmented at organisational boundaries

Digi Yatra has simplified parts of the traveller’s journey, but the interview identifies a structural limit. Entry and security, airline check-in and boarding, and immigration may operate under different identity and legal regimes. A traveller may have to prove identity again not because face matching is impossible, but because one agency’s permission to process data does not extend to another. Ankalgi contrasts this fragmentation with more continuous biometric experiences in parts of Asia. Interoperability is therefore as much a problem of standards, trust, lawful purpose and governance as facial-recognition accuracy. The editorial test is not simply whether a face is recognised. It is whether the user can move efficiently without creating a single uncontrolled store of sensitive biometric information.

The missing ingredient is a shared operating rulebook

Near the close, Ankalgi calls for a unified aviation data-exchange policy and proposes reducing asset idle time relative to passenger movement as a leadership metric. These ideas are connected. An airport cannot optimise resources across firms when everyone defines events differently, restricts access to operational data or refuses accountability for a bad outcome. A consistent event model—aircraft on-block, bags available, stand released, security queues forecast—would make useful comparisons possible. But a central exchange should not mean unlimited disclosure: cybersecurity, competition and privacy boundaries still apply. The most significant outcome would not be more data. It would be a common ability to identify a capacity constraint, assign an owner and test whether a response worked.