Travel brands are using AI agents in advertising to answer campaign questions, build and launch campaigns from a brief, move budget between channels and adapt creative, while hotel booking itself is moving inside AI assistants where ads now appear. In almost every documented case from 2025 and 2026, the agent recommends or drafts, and a person, a written objective or a permission set decides how far it can act.

TL;DR: The eight examples below come from Red Roof, Navigator, Amadeus, Expedia Group, Accor, Koddi, Iberia and Google's AI Mode, each sourced to the company itself or to trade press. Most travel advertising agents today sit at the recommend or draft stage, with a few acting automatically inside limits someone else has set. The harder question for travel marketers is not what an agent can do, but who can check what it did.

Why is travel advertising a hard test for AI agents?

Travel inventory is perishable: a hotel room or airline seat left unsold on the night cannot be sold later, so campaigns chase demand that shifts by season, route and day. An advertising agent that misreads a brief, or keeps spending after a booking window has closed, does damage quickly and visibly.

Testing an agent before it spends is one answer to that risk. As a worked example, a hotel group could use the Alkimi Gym to build its own buying agent, set the rules it works to (no spend against stay dates outside the agreed window, for instance) and spar it against seller agents before any real money moves.

Which travel and hospitality companies are using AI agents in advertising?

Each example records only what the company or the trade press said the agent does, and no performance figures are repeated.

1. How does Red Roof use a voice agent for campaign decisions?

In March 2026, AdExchanger reported that the hotel chain Red Roof took part in the beta of Athena, a voice-activated AI agent from Zeta that answers natural-language questions with campaign statistics and suggests revisions to campaigns in flight. Red Roof president Zack Gharib said the agent lets his team guide campaign creation, deployment timing and audience development with simple verbal commands.

Zeta chief executive David Steinberg said Athena brings steps that lived in different parts of the platform into a single workflow "without all the back and forth." He described it as a "co-pilot" that offers guidance on spend and deployment, which leaves a person between the suggestion and the spend.

2. How is a travel media network testing agentic media buying?

In April 2026, AdExchanger reported that Kargo had released Project Kera, a chat-based planning and buying environment tuned for media buyers, in closed beta. The publication named the travel media network Navigator among the testers running full campaigns, alongside Hershey and the independent agency Wpromote.

Kargo chief executive Harry Kargman told AdExchanger the aim is to "aggregate all biddable and non-biddable inventory into a single platform."

3. What does the Amadeus Travel Advertising Platform automate?

Amadeus introduced its Travel Advertising Platform in May 2026 at its Advertising Summit in Nice, drawing on Accenture's agentic framework for campaign execution. Its first live capability, the Omnichannel Budget Allocator, uses AI to evaluate campaign performance and automatically reallocate budgets across channels and campaigns against defined objectives, and is already used by Amadeus Advertising teams within managed services.

Amadeus says the automation "is designed to enhance human strategy, ensuring decisions remain transparent, controllable and outcome-driven." Dan Ciocoiu, head of advertising solutions at Amadeus, described the platform as "using AI to support better decisions and connect demand insight to coordinated action across channels."

4. What AI agents did Expedia Group announce for partners at Explore 2026?

At its Explore 2026 partner conference in Las Vegas, Expedia Group announced three agents for its Partner Central platform. Partner Companion identifies performance issues and commercial opportunities, a Content Agent detects what information travellers still need from a listing, and Autonomous Distribution pre-fills listings from trusted sources.

For marketing, Expedia Group described "business-to-agent" (B2A) capabilities, including OpenAI Ads and Trip Matching with Meta, to help partners be found by the AI surfaces where trips are planned. The company also said its Responsible AI council reviews all high-risk AI deployments before they scale.

5. How does Accor's agency use an AI agent for paid search?

A Skai case study published in July 2026 describes Zenith International managing paid search for Accor across two large accounts on Google and Microsoft Ads. Zenith deployed Skai's Celeste AI across both and wrote structured prompts that standardised how it retrieved, analysed and compared data, so the output was consistent whoever ran the query.

The team compared the agent's analysis with platform search query reports before relying on it. Each query then returned optimisation recommendations rather than a spreadsheet, which keeps the agent at the recommend stage with a check built in.

6. How do travel marketplaces let agents act on their ad platforms?

Koddi, whose ad technology serves travel among other sectors, described Koddi Ads MCP in March 2026. It exposes the Koddi Ads platform to agents through an open Model Context Protocol server, "with the same permissions and role-based controls that govern the Koddi Ads platform."

Koddi's own travel example is a marketplace that automatically traces performance shifts to changes in bid multipliers tied to seasonality or inventory constraints. The agent can read and write to the platform, but only within permissions a human user already holds.

7. How does Iberia use agentic AI for ad creative?

In June 2025, AdExchanger reported that the Spanish airline Iberia had brought media in-house using Flight Control, a self-serve tool from the AI personalisation platform Clinch. One Iberia ad can be broken down and rebuilt into versions for each audience segment and media channel.

Clinch chief executive Oz Etzioni said the tool's agentic AI analyses each element within an ad to determine "what resonates best with who and why." Carlos Fernández Suárez, Iberia's digital marketing manager, said the ability to make real-time changes to ads is "critical in our industry."

8. What happens when the AI assistant itself becomes the ad placement?

Here the traveller's assistant is the agent. In July 2025, Koddi reported that Gemini-powered Google search had started showing paid, feed-based hotel listings inside AI-generated itineraries, so a search for a London itinerary can lead to hotel ads at the "where to stay" step.

In August 2026, Google introduced hotel booking in AI Mode in the US, with partners including Booking.com, Choice Hotels International, Expedia, Hilton, IHG Hotels & Resorts, Marriott International and Wyndham Hotels & Resorts.

For travel advertisers, this shifts part of the job from buying impressions to supplying accurate, structured feeds and content that an assistant can read and act on.

What do the eight examples have in common?

Mapped against the five stages of earned autonomy (observe, recommend, draft, human-approved action and bounded automatic action), most of these agents sit in the middle. Red Roof's co-pilot, Zenith's work for Accor and Expedia Group's partner agents observe and recommend, while Iberia's creative tool and Navigator's test campaigns draft and execute. Only the Amadeus allocator and agents working through Koddi's server describe bounded automatic action, against defined objectives or existing permissions.

In every case the limit is written down outside the agent: an objective, a role-based permission, a review council or a set of structured prompts. The public descriptions focus on the advertiser's side of the work and say less about how the other party to a buy would see the same record of what an agent agreed.

What should a travel marketer check before giving an agent more control?

Which earned autonomy stage the agent works at today, and who signs off before it moves up a stage.
Where its limits are written down, for example in a media buying mandate, an objective or a permission set.
Whether its outputs have been checked against an independent source, as Zenith did with search query reports.
How it handles date-bound inventory, such as a stay window or a route that closes.
Whether both sides of a deal can retrieve the same record of what was agreed.

Where do standards such as AAMP and AdCP fit?

Most of the examples above connect an agent to a single platform. Once buying agents deal with seller agents at other companies, shared standards matter more.

AAMP, IAB Tech Lab's Agentic Advertising Management Protocols, and AdCP, the Ad Context Protocol published by AgenticAdvertising.org, are both written for agent-based advertising. A travel marketer assessing an agent can reasonably ask which of them it works with.

Where does this leave travel advertisers?

Travel marketers already use agents for the slow, repetitive parts of advertising. The next step, letting them act on bookable, date-bound inventory, needs evidence of how an agent behaves before it spends.

Concourse, the live benchmarking product from Alkimi for AI media-buying agents, found in Bench v1 that 95 of 110 signed contracts promised a viewability floor below 70%. A completed buy and a good buy are different things, so the sensible route for travel brands is to test agents against counterparties first and keep human approval at the deal level until the record shows they have earned more.

Frequently asked questions

What is an AI agent in travel advertising?

An AI agent in travel advertising is software that pursues an advertising goal, such as filling rooms for a set of dates, by analysing performance, drafting campaigns or moving budget. It differs from older automation because it chooses its next step itself, within limits a person sets.

Are travel brands letting AI agents buy media on their own?

A few examples come close: the Amadeus Omnichannel Budget Allocator reallocates budgets automatically against defined objectives, and Navigator has tested full campaigns in Kargo's Project Kera. Most other documented cases keep the agent at the recommend or draft stage.

How do AI assistants change hotel advertising?

Paid hotel listings now appear inside AI-generated travel plans in Google search, and AI Mode in the US lets travellers book hotels with partners such as Hilton, Marriott International and IHG Hotels & Resorts. That puts more weight on accurate feeds, availability and content an assistant can read.

What should travel marketers ask an AI agent vendor?

Ask which earned autonomy stage the agent works at, where its limits are written and who approves anything beyond them. Ask too whether its outputs can be checked against an independent record, and which standards, such as AAMP or AdCP, it works with.

AI Agents in Travel and Hospitality Advertising: 8 Sourced Examples