Agentic commerce is reshaping how travelers discover flights and redeem loyalty points. Unlike retail, AI-driven discovery often strengthens intermediaries and leaves airlines further from the customer interaction while still relying on airline systems to generate the answers.
KasadaIQ
TL;DR
AI-driven discovery is shifting upstream. Travelers increasingly ask AI assistants for flight options, and those systems often rely on OTAs and aggregators rather than interacting directly with airlines.
AI amplifies the intermediary layer in travel. Unlike retail, where AI can connect customers directly to brands, the travel distribution ecosystem means airlines can end up further removed from the discovery interaction.
Airline infrastructure absorbs much of the cost. Automated search, pricing monitoring, and loyalty optimization tools generate large volumes of queries against airline systems, often without producing bookings.
The challenge is no longer just bots. Airlines now face a spectrum of automated actors, from legitimate AI customer agents to high-frequency pricing monitors and malicious automation.
The discovery inversion
AI is changing how travelers discover flights and redeem loyalty points. In travel, the implications are very different from most industries.
In many sectors, AI pushes customers closer to the brand. Ask an AI assistant to find a pair of Nike shoes and it can often reach Nike directly or connect to a tightly controlled retail channel.
Travel works differently.
When a traveler asks an AI assistant for the cheapest flight from Sydney to Singapore next month, the assistant usually does not go directly to an airline. Instead it pulls information from aggregators, metasearch platforms, and OTAs. These intermediaries sit between the traveler and the airline's inventory.
The airline still supplies the product. But it has moved further from the conversation.
Discovery is no longer happening at the airline’s front door.
The interaction increasingly looks like this:
Traveler → AI assistant → OTA / metasearch / aggregator → airline inventory
That dynamic is likely to define agent-mediated travel discovery. AI amplifies intermediaries rather than direct brands.
Why airlines cannot pull a Nike
Airlines do not have the luxury of behaving like a direct-to-consumer retailer.
They operate inside one of the most complex distribution ecosystems in any industry. Price parity expectations, full-content agreements, indirect sales channels, and decades of GDS infrastructure shape how airline products are sold.
Because of that structure, airlines cannot simply reserve a better price for their own website the way a direct brand might. Instead, they compete through bundling, ancillaries, loyalty benefits, personalization, and cost of sale.
Personalization depends on understanding the customer.
A business traveler may value flexibility and priority services. A family is more likely to care about baggage allowances, seating, and overall trip cost. When airlines know who the traveler is, they can shape the offer accordingly.
When discovery happens through an AI layer sitting above OTAs and aggregators, that context disappears.
In retail, AI pulls customers toward the brand.In travel, it often pushes them toward intermediaries.
As AI-driven discovery becomes more common, airlines risk being reduced to inventory providers inside someone else's interface.
The infrastructure cost nobody is talking about
Flight search is computationally expensive. Generating a priced itinerary requires calculations across schedules, availability, fare rules, ancillaries, and distribution logic.
Historically, airlines managed this through the relationship between searches and bookings, often referred to as the look-to-book ratio.
That ratio has grown dramatically over time.
Early internet travel produced hundreds of searches per booking
Metasearch increased this to thousands
Modern travel platforms can generate tens of thousands or more
AI search contributes to this trend, but it is not the only factor.
Dynamic pricing has changed the incentive structure.
When prices change frequently, monitoring those changes becomes valuable. Pricing analytics and loyalty optimization services query airline systems repeatedly to map fare behavior, detect inventory changes, and identify cheaper combinations.
In practice, they are attempting to reverse-engineer airline pricing logic.
AI agents accelerate this dynamic. They can generate queries automatically, refine results, and explore alternative scenarios at scale.
Dynamic pricing created the incentive. Automation created the capability. AI agents add the intelligence.
Each query places load on airline systems even when no booking follows.
Loyalty is becoming an optimization surface
This dynamic becomes especially visible in loyalty programs.
Tools such as seats.aero allow travelers to search reward seat availability across multiple airlines using automated queries. For users, these tools provide enormous value. They surface rare award seats, unusual routing opportunities, and redemption options that would be almost impossible to find manually.
At the same time they introduce a new form of information asymmetry.
Travelers using automated discovery tools can identify the most efficient redemption opportunities across multiple airline programs simultaneously.
Travelers with automated search tools see opportunities that others never knew existed.
For airlines, this creates two challenges.
First, these platforms rely on high-frequency automated queries to map reward availability and pricing behavior. Airlines effectively subsidize the infrastructure load that powers someone else's optimization layer.
Second, the loyalty relationship itself begins to shift.
The airline still owns the points and the seat inventory. But the discovery layer, and increasingly the perception of where value exists, may sit elsewhere.
Tools like seats.aero are likely only the beginning.
AI agents could monitor reward availability across multiple programs continuously, analyze pricing behavior, and surface optimal redemption opportunities in real time. The AI becomes a personal loyalty optimizer, scanning airline systems on behalf of the traveler.
The question for airlines is not whether this will happen.
The real question is this:
What happens when thousands of AI agents are simultaneously interrogating reward availability, pricing behavior, and fare conditions across your network?
The agentic internet is a governance problem
As AI agents become more capable, the conversation shifts from what they can do to who controls access.
Recent disputes involving autonomous AI browsing and major online platforms are already testing the boundaries of automated access to digital infrastructure.
The direction of travel is becoming clearer. Platforms will retain the right to control how automated systems interact with their systems, even when those systems claim to act on behalf of users.
For travel, this matters because agentic access is unlikely to remain completely open. It will more likely evolve toward:
structured APIs
trusted partner channels
rate-limited access
selective distribution relationships
This is already familiar territory for airlines.
The harder problem is that not all automated traffic behaves the same way, and automated systems do not always identify themselves honestly.
Airline platforms may now receive traffic from:
AI-assisted customers generating genuine demand
AI retrieval agents building search results
automated pricing monitors
loyalty optimization tools
model training crawlers
malicious bots spoofing AI user agents
Some of this traffic creates value. Some extracts it. Some is harmful.
The old question was simple.
Human or bot?
That question is no longer sufficient.
The more useful question today is:
Which agents are interacting with our systems, and what should we do about them?
From bot mitigation to Agentic Trust
For years the playbook was straightforward. Identify malicious bots and block them.
That approach no longer reflects reality.
Some automated traffic now represents legitimate customer interaction. Some drives bookings. Some creates infrastructure cost without commercial value. Some distorts analytics or pricing signals.
The challenge is no longer simply stopping bad actors.
It is understanding and governing the full spectrum of AI agents interacting with airline systems.
This is where the concept of Agentic Trust begins to matter.
Agentic Trust focuses on answering practical questions such as:
Which AI-driven channels generate real demand
Which agents are querying pricing or loyalty systems without commercial intent
Which traffic should be allowed, throttled, redirected, or blocked
How performance and fairness can be preserved as AI becomes embedded in travel discovery
Visibility is the starting point
Travel has always been shaped by technology, distribution, and the battle for customer relationships.
Agentic AI introduces another layer into that ecosystem, and in many cases, it strengthens the role of intermediaries.
Travelers may increasingly rely on AI assistants to discover and compare flights before they ever reach an airline website. Those AI systems often sit on top of existing aggregators and distribution channels.
The result is a strange dynamic.
Airlines provide the product. Airlines run the infrastructure. But the discovery experience may happen somewhere else.
The airlines that succeed in this environment will not simply be those deploying AI internally.
They will be the airlines that develop clear visibility into the automated systems interacting with their infrastructure, understand which interactions create value, and manage them accordingly.
It begins with one simple capability.
Knowing who, or what, is actually at the door. Don't know? Find out.
Read more
Defeats threats not your customersDefeatsthreatsnotyourcustomers
