AVGEEK
The software running behind the flight
Somit Goyal of IBS Software examines legacy systems, airline retailing and the transition to cloud-based operations.
The conversation
Every time you book a flight, check in online, or reserve an in-flight sleeping pod, there is a high-tech engine working silently behind the curtain. In Episode 78 of The Indian AvGeek , host Vishal Mehra sits down with Somit Goyal , CEO of IBS Software.
Bringing an enterprise tech background from Microsoft, SAP, and Oracle, Somit shares a frank assessment of aviation's legacy technology debt, why NDC is merely an infrastructure "plumbing fix," and how cloud-native, AI-first architectures are transforming commercial and flight operations globally.
In this episode, they dive into real-world case studies:
Korean Air: Migrating live flight operations ("open-heart surgery while running a marathon"). Air New Zealand: Powering ground-breaking Skynest sleep pods using iRetail. Indian Carriers: Shifting from treating IT as a cost center to leveraging software as a strategic differentiator.
Key Takeaways
- The publisher frames the Goyal episode around aviation’s legacy technology and the systems behind the passenger interface. The editorial issue is how reservations, flight operations, crew and commercial systems work together; complexity increases the need for disciplined modernisation rather than excusing fragmented architecture.
- NDC improves distribution communication; it is not the whole retailing business. IATA describes it as a data-exchange format based on Offer and Order processes. An airline still needs the commercial decisions and connected capabilities for inventory, payments, order servicing and disruption handling; this distinction is analysis, not a verified quotation from Goyal.
- Korean Air is a documented iFlight deployment example. IBS describes a phased move from Operations Control to Crew Tracking and Crew Planning. That vendor account supports the migration sequence, not independently measured performance gains; continuity, training and rollback arrangements are editorial tests, not safeguards individually verified in this project.
- AI can assist crew and disruption decisions only within verified operational and regulatory constraints. Duty-time limits, qualifications and accountable approval cannot be replaced by a persuasive interface. This is an editorial assessment of responsible deployment, not a claim that the episode demonstrated autonomous dispatch or a measured saving.
- Skynest illustrates inventory sold by time window rather than simply by flight-long seat allocation. IBS’s 2 June 2026 announcement describes iRetail support for four-hour sessions and a launch planned for later that year. It verifies the announced product-support role, not actual service commencement, customer take-up or profitability.
- The Indian-carrier theme is software as operating capability rather than merely an IT expense. At increasing fleet scale, integration and change-management costs should be judged against measurable reliability, service and total cost of ownership; buying more software is not itself a strategy.
- The evidence has limits: the supplied edition relies on publisher topic and chapter material, not an independently audio-verified transcript. Corporate announcements corroborate named examples, while AI prescriptions and commercial implications are editorial inference. No exact guest quotation, complete recording coverage or independent vendor-performance audit is claimed.
Additional subjects
Editorial analysis
Editorial context
Airline digital transformation is often sold as a retail experience. The more consequential work may be migrating flight operations, crew planning and inventory systems without interrupting the business.
The architecture beneath the customer promise
A traveller sees a booking screen or boarding pass, not the reservation, operations and crew systems behind it. The publisher’s description places legacy technology, Korean Air, Skynest and Indian carriers among this episode’s topics. That establishes the editorial subject, not the precise order or wording of Goyal’s remarks. The analytical question is how systems remain coordinated when schedules change. Safety-critical operations, time-sensitive inventory and compliance create genuine constraints, but do not remove the value of modular interfaces, observable data and tested change processes. These are editorial criteria for modernisation, not attributed quotations.
NDC is not a retail business model
IATA defines NDC as a data-exchange format based on Offer and Order management processes, enabling airlines to distribute offers across channels. It is therefore more precise than calling it only a booking interface, but it is not the complete retailing operating model. Merchandising, payments, inventory, order servicing, accounting and recovery from disruption still require connected capabilities and commercial decisions. The publisher uses a plumbing analogy; that wording is not independently verified here as Goyal’s exact speech. The editorial distinction is between improving communication and delivering the whole customer proposition.
The hardest migration is the one nobody notices
Korean Air is the publisher’s named migration case. IBS’s public announcement separately describes Operations Control deployment followed by Crew Tracking and Crew Planning. The article’s header and body give conflicting years, so this edition does not assign a precise go-live date. The documented sequence is a vendor account, not an independent measure of operating improvement. Continuity, user training, data quality and rollback planning are appropriate editorial tests for any mission-critical migration; the announcement does not independently verify each safeguard in this project. A serious assessment needs airline-side performance and reliability measures rather than an inference from a go-live announcement alone.
What AI can really do in operations
AI-assisted planning is an editorial application of the episode’s software theme, not a recovered verbatim account of a particular demonstration. Crew and disruption decisions propagate through aircraft rotations, passenger connections and recovery costs. An assistant may help identify options, but duty-time rules, qualifications, agreements, operational status and accountable approval still have to be checked. A conversational interface does not validate those constraints. The test is measurable improvement with auditability and appropriate human responsibility. This edition makes no quantified disruption-cost claim, verified productivity promise or assertion of autonomous dispatch.
When a product defeats the seat map
Air New Zealand’s Skynest is the publisher’s named iRetail example. In its 2 June 2026 announcement, IBS described support for four-hour sleeping sessions and a commercial debut planned for later that year. A time-window product raises a different inventory problem from a seat allocated for an entire flight; IBS names Product Catalogue and Stock Keeper among the supporting capabilities. The wider implications for eligibility, disruption handling and product design are editorial analysis. An announced supplier role does not verify that passenger service has begun, how many customers have bought it or whether it has increased yields or profit.
Scale changes the CIO’s job
The publisher identifies Indian airlines’ shift from treating IT as a cost centre towards a strategic differentiator. The editorial implication is that operating complexity can make fragmented workarounds costly, but buying more software is not itself a strategy. Reliability, service, integration, deployment risk and total cost of ownership provide a more useful scorecard. Smaller carriers also need to consider fixed implementation costs. A subscription changes the timing of expenditure, not the need to demonstrate value. These are criteria for evaluating investment, not verified claims about the systems or budgets of particular Indian carriers.
The economics of invisible reliability
Reliable operations require more than a fashionable interface or architecture label. Cloud-native platforms and AI are propositions to test against recoverability, cyber resilience, service levels, integration economics and migration risk. This edition’s synthesis is that software becomes strategic when it helps make airline promises dependable. It is not an independently verified statement of Goyal’s precise thesis, a certification of one supplier or a finding that every new platform outperforms an established system.
Questions & answers
What is the episode’s central proposition?
Modern aviation software should support connected, measurable airline operations rather than simply replace old interfaces.
Why is NDC not enough?
Distribution messaging alone does not solve merchandising, payments, order servicing, inventory and operational integration.
What makes the Korean Air case significant?
A phased deployment of mission-critical flight-operations and crew systems was publicly announced; independent outcome data still matter.
Can agentic AI dispatch flight crews?
It may assist planning, but legal, qualification, operational and safety constraints require independently verifiable checks.
Why does Skynest matter beyond a novelty product?
A bookable time window introduces inventory logic that traditional flight-long seat allocations do not naturally represent.
Sources & attribution
Original recording: The Indian AvGeek — Episode 78. First released 7 September 2026.
Publisher provenance: The supplied editorial package identifies publisher chapter notes as a source. This review checked the saved publisher description and supplied archive material, but did not independently retrieve the complete timestamped chapter list, verify a full transcript or align the released audio. Topic notes establish subjects, not exact statements or complete recording coverage.
Standards context: IATA — New Distribution Capability explains the Offer-and-Order-based data-exchange format.
Corporate examples: IBS — Korean Air iFlight deployment describes a phased deployment; its header and body contain inconsistent years, so no precise go-live date is assigned here. IBS — iRetail and Skynest, 2 June 2026 documents announced product support and a then-planned launch, not independently measured airline outcomes.
Editorial interpretation: The five Q&As are explanatory paraphrases, not direct guest quotations. AI safeguards, inventory implications and investment criteria are editorial analysis; vendor announcements do not independently establish performance, take-up or profitability.