Six decades after Sabre launched the world’s first computer reservation system, the travel industry continues to treat early computational breakthroughs as permanent foundations rather than stepping stones. While user interfaces have evolved from mainframe terminals to mobile apps and basic chatbots, the underlying operational architecture across many organizations remains untouched.
This creates an operational paradox. AI travel agents can cross-reference thousands of data points to assemble a complex international itinerary in seconds. Yet, changing that same booking often requires hours of call-center manual labor and dozens of disparate screen clicks.
Deploying artificial intelligence over century-old system architectures and organizational structures limits potential returns. To unlock genuine business value, travel enterprises must change focus from marginal cost-cutting to structural reinvention.
The Limits of “So-So Automation”
Most travel organizations suffer from what Nobel laureate Daron Acemoglu terms “so-so automation”: technology that substitutes human labor without generating novel value. Applying AI merely to replace contact center agents quickly caps financial benefits. Conversely, directing AI toward value creation, such as embedding intelligent ancillary offers into real-time service conversations, yields scalable, compounding returns.
When organizations bolt new tools onto legacy processes without changing outcomes, they fall into “transformation theater”. An internal assessment across a five-level maturity scale highlights this reality:
- Fragmented: Scattered, uncoordinated departmental experiments with no clear accountability.
- Proven: Isolated pilots yielding early ROI and initial user evidence.
- Operational: Core workflows deployed into production under standard controls.
- Integrated: Shared enterprise rails for data, governance, and reusable components.
- Compounding: A self-funding portfolio engine where initial gains finance subsequent innovation.
Benchmark data across Globant’s 30+ major travel and airline engagements reveals that the overwhelming majority of the industry remains stuck at Level 1 or 2. High spending yields low maturity. The primary issue is not an AI shortage, but an enterprise consistency failure caused by disconnected tools operating in silos.
The Three-Layer Architecture of Transformation
To break through the automation plateau, digital transformation must address three interconnected layers. Read from the top down, they outline the strategic agenda; read from the bottom up, they reveal the true order of dependency: people (WHO) enable capabilities (HOW), and capabilities enable experiences (WHAT).
- WHAT (The Experience Layer): This layer features two distinct faces because today’s travel consumers are both humans and machines:
- Opportunity 1: The Human Experience (Generative UI): Rather than forcing every customer through a static menu or app interface designed years ago by a committee, generative AI in travel constructs the ideal screen in real time, tailored specifically to the user’s immediate context. For example, a family planning a trip gets a visual, collaborative canvas, while an executive in transit gets three flight options and a single confirm button.
- Opportunity 2: The Machine Customer (AI Agents): Instead of building only for human eyes, enterprises must open their systems to autonomous AI agents that read APIs rather than navigate visual branding. Because static booking funnels and date pickers currently block these agents, companies must adopt open protocols such as MCP or UCP so that external AI can directly query inventory and execute purchases.
- HOW (The Capability Layer): The main bottleneck to transformation in travel is structural, as resolving a single canceled flight may require coordinating data across many siloed systems, like Reservations, Loyalty, Baggage, Crew, and Airport operations, each guarded by its own owner and budget. No AI will orchestrate what the organization itself doesn’t allow to be orchestrated. Overcoming this requires moving beyond front-end screens and dismantling internal fiefdoms, allowing AI to connect end-to-end capabilities through modular “power outlets” built on decoupled services, independent business logic, and clean APIs.
- WHO (The People Layer): Execution steps can be delegated to automated systems, but accountability cannot. Internal silos and fiefdoms do not fall to APIs; they fall to redesigned roles, updated incentives, and clear accountability structures. As repetitive tasks shift to machines, existing staff move from executing manual steps to framing problems, orchestrating workflows, and judging quality. At the same time, new roles emerge to curate data and set decision boundaries for AI, while heavy supervisory layers flatten into cross-functional teams that own business capabilities end-to-end.
Executing via “Value Cells”: Redrawing the Blueprint
Broad, multi-year programs frequently die in committee because they attempt to renovate the entire house at once. The practical path to enterprise transformation works in the opposite direction, relying on self-funding deployment units called “Value Cells”:
- Prove: Isolate one bounded business workflow, deploying it to real users over 10 to 12 weeks with measured ROI.
- Scale: Connect three or four proven cells over shared rails, such as data, governance, and reusable components.
- Transform: Let each cell fund the one after it, aiming at new revenue rather than just margin cost-cutting.
This cell-by-cell model prevents legacy inertia. Incumbents rarely lose because challengers are smarter; they lose because new entrants start from a blank page while established companies spend resources defending legacy systems. Without technical debt or internal fiefdoms, a startup can launch an agent-ready travel business in eighteen months. While incumbents retain superior inventory, trust, and operational depth, challengers win on reachability, and an unreachable advantage provides zero value to a machine customer.
From Legacy Inertia to Compounding Capability
An AI-first organization operates as a self-financing engine powered by real-time generated interfaces, modular capabilities, and outcome-accountable teams. Instead of defending the floor plan inherited from the last century, the path forward for travel enterprises is redrawing operations one cell at a time. Sixty-two years ago, the travel industry bet its business on computation and changed global commerce forever. The opportunity today is to do it again and build a foundation designed to evolve.
Ready to move beyond legacy limits? At Globant’s Airlines AI Studio, we combine deep travel domain expertise with AI-native platforms to help global carriers dismantle legacy silos and scale agentic commerce capabilities. Connect with our experts to redraw your operational blueprint.
