Thinner Interfaces, Thicker Trust: What Airlines Can Learn from AI Design

The Interface Is Thinning—But the Experience Is Getting Heavier

Walk onto any modern aircraft today and you'll see the same paradox. The physical cockpit is shrinking, the screens are consolidating, and the flight crew's tablets are replacing stacks of manuals. But the actual job of flying—and the job of running an airline—is getting more complex, not less.

This mirrors something happening across product design. As AI starts to act on our behalf, the visible interface gets thinner while the invisible rules multiply. For airlines, this shift is happening at every layer: from how passengers book tickets to how pilots handle weather diversions.

From "Man Looks for Function" to "AI Understands Intent"

Traditional airline software assumed a simple premise: the human must learn the system. You needed to know that the fare code is on the third tab, that the rebooking button lives in the lower right corner, that the "remarks" field is where you type the waiver code. A huge amount of design work went into reducing that learning curve—fewer clicks, clearer labels, better layouts.

AI flips this. Instead of the passenger understanding the airline's booking engine, the engine tries to understand the passenger. You type "I missed my connection, get me on the next flight to Chicago"—and the system figures out which tools to call, what rules apply, and how to make it happen.

This is a shift from flow design to intent design. The old question was: where does the user go next? The new question is: what did the AI actually understand? And that raises a fresh cost that designers must minimize—not just the cost of clicking, but the cost of being misunderstood by a machine.

Fewer Pages, More Rules

It's tempting to think that fewer screens mean simpler design. Not true. In an AI-driven airline system, the hard problems are not about button placement. They're about behavior:

• When should the AI act automatically—like rebooking a missed connection—and when should it ask first?
• What happens if the AI makes a wrong assumption and books an overnight stay at the wrong hotel?
• Can the passenger see what the AI did? Can they undo it?
• Who is in control when the AI hits a situation it can't handle—like an airport closure?

These rules rarely appear on a static page. They live in the system's logic. But they shape the experience far more than any visual design. The interface gets thinner, but the experience gets thicker.

From Usability to Delegability

For decades, airline websites have been obsessed with usability. Can the user find the flight? Is the checkout fast? Does the mobile app work on a shaky airport Wi-Fi connection? That's all still important. But when AI starts performing actions on behalf of passengers—rebooking, refunding, rerouting—a new metric emerges: delegability.

Delegability is the answer to a simple question: Do I trust this system enough to let it act for me? A brilliant AI that never makes mistakes but keeps no record of its actions is not delegable. A fast AI that silently changes my seat without asking is not delegable. The goal shifts from "make it easy to use" to "make it safe to hand over."

This is not just about passenger-facing tools. It applies to airline operations as well. Dispatchers feed weather and fuel data into AI that suggests routing changes. Maintenance planners rely on AI to predict part failures. In each case, the human's trust in the machine determines how much authority the machine gets.

Sometimes the Right Move Is to Ask One More Question

Traditional UX preached brevity. Fewer clicks, fewer steps, fewer confirmations—that was efficiency. But in an AI-powered airline system, that principle breaks down.

Imagine a passenger says, "Cancel my flight." The AI could immediately process the cancellation, wave the fee, and send a refund confirmation. But what if the passenger meant "cancel the flight I booked for my boss, but keep mine"? Or what if they meant "change the date" but used the wrong verb? An immediate action might be efficient but wrong.

Good AI in this context knows when to pause. It asks a clarifying question not because it's stupid, but because it cares about getting it right. This is what I call boundary design—designing not just what the AI can do, but what it should do, and when it must stop and hand control back to a human.

Designing Behavior, Not Just Pages

If traditional interface design is like building a stage—entrances, pathways, lighting—then AI experience design is more like directing a play. You're deciding when the AI speaks, when it stays silent, when it makes a suggestion, when it takes action, and when it openly admits uncertainty.

This is AI behavior design, and it's becoming a core competency for airlines. Consider the difference between two customer-service bots:

Bot A answers questions with confidence, but never reveals its uncertainty. When it can't find a booking, it says "I couldn't find your reservation" and ends the chat.

Bot B says, "I'm not sure I found the right booking. There are two reservations under your name—one from last week and one for tomorrow. Which one did you mean?"

Bot B is doing behavior design. It's managing expectations, acknowledging ambiguity, and offering a path forward. That's the difference between a system that feels smart and one that feels trustworthy.

Designing Expectations: The Passenger Needs to Know What Happens Next

With traditional software, behavior is predictable. Click "Download boarding pass" and you know what you'll get. With AI, the passenger can't always predict what will happen. Will the AI just suggest a new flight, or will it actually book it? Will it change my seat, my meal, my luggage allowance?

This is where expectation design comes in. The AI doesn't need to explain everything, but it must set accurate expectations before acting. For example, before rebooking a missed connection, the AI should say: "I'll check alternative flights and rebook you on the earliest one. I'll send you the new boarding pass. Does that work?"

After acting, it should confirm what it did: "You've been rebooked on flight 442, departing at 6:30 PM. Your original seat preference has been applied." This closes the loop and builds confidence.

Reversibility: The Underrated Feature

Why are passengers hesitant to let AI handle things? Often, it's not because they think the AI is dumb. It's because they fear they can't undo its mistakes. This is where reversibility becomes critical.

For an airline, reversibility means:

• Every automatic action has a log.
• Every change can be undone.
• Every confirmation is optional.
• Every process can be interrupted.
• Every failure has a human fallback.

These aren't flashy features. They don't make headlines. But they determine whether passengers actually let the AI take over. A trustworthy AI isn't just one that does things well—it's one that lets people change their minds.

From UI Standards to Experience Governance

In the past, airlines maintained brand consistency through design systems—unified colors, typography, component libraries. That's still important. But as AI spreads across the airline's digital ecosystem, a new kind of consistency is needed.

Does every AI touchpoint—booking, check-in, rebooking, baggage tracking—use the same confirmation mechanism? Do different AI skills have clear permission boundaries? Is there a consistent way to escalate to a human? Can every action be verified, traced, and reversed?

This is experience governance. It's not about standardizing pixels; it's about standardizing how intelligent systems interact with people. Airlines that master this will create a coherent, trustworthy AI experience. Those that don't will end up with a patchwork of clever but chaotic tools.

Design Value Hasn't Disappeared—It's Migrated

AI will certainly reduce some traditional design work. Standard pages, repetitive layouts, even basic front-end code—these are increasingly automated. But the real question isn't "how many design jobs will survive?" It's "where do the new experience problems live?"

For airlines, the answer is clear: the new problems live in intent, behavior, boundaries, and trust. The industry doesn't need fewer designers; it needs designers who can shape how AI behaves, not just how it looks.

Shaping Trustworthy Intelligence

If design is just about making things look good, AI is indeed eating away at that work. But if design is about deliberately shaping the relationship between people and systems, then AI is expanding the job, not shrinking it.

Airlines have always been in the trust business. Passengers trust pilots with their lives, trust mechanics with the airframe, trust dispatchers with the route. Now, they're being asked to trust algorithms with their itineraries, their refunds, their time. That trust won't come from a beautiful interface. It will come from systems that understand intent, respect boundaries, set expectations, and offer a way back.

The next great airline design challenge isn't a better booking flow. It's designing an AI that deserves to be trusted.