Travel industry confronts hidden costs of AI deployment
Travel companies are discovering that artificial intelligence, sold as a cost-cutting solution, carries substantial hidden expenses that many organizations underestimated during pilots and early tests.
The problem surfaced at recent industry conferences. Riccardo Vittoria, CEO of AI platform Acai, told attendees at the Airline Distribution 2026 conference in March that while agentic AI can perform tasks faster than humans, "it's still expensive." Similar concerns emerged from hospitality leaders at the same time.
The cost structure breaks down into several components. Infrastructure licensing, data center electricity, and specialized talent all carry price tags. Energy costs are rising due to global market disruptions. A Gartner report predicts more than 40% of agentic AI projects will be canceled by the end of 2027 due to escalating costs.
The math depends on what you're replacing
Economics shift dramatically depending on the use case. When AI replaces human labor-such as handling flight changes or rebooking disruptions-the savings are substantial. An AI-powered rebooking that costs a few dollars can replace a human agent handling the same task for $30 to $50.
The calculation inverts when AI replaces existing technology systems. An AI agent handling a new booking through a call center could cost 80 cents to $10, depending on complexity. The same ticket booked online costs just a few cents.
Vittoria warned that airlines, travel management companies, and online travel agencies are "rushing to build agentic booking experiences without understanding the unit economics." The risk is deploying expensive solutions where cheaper alternatives already exist.
What HRS learned from real deployment
Accommodation platform HRS invested in AI over two years and launched an AI-powered Copilot platform. Chief product officer Martin Biermann said the costs are real: infrastructure licensing, RFP management, contracting, auditing, and regional staff to support the system.
HRS found the biggest gains in speed. AI-driven optimization cycles increased from roughly 10 per year to 40-50, generating an additional 10% yield. The company reduced headcount requirements by 95%-work that previously required four employees now requires one-fifth of one employee.
That efficiency comes at a price. AI-related talent costs are rising roughly 30% annually, and specialized roles take 50 times longer to recruit than standard positions.
Enterprise infrastructure gives larger players an edge
Wyndham Hotels & Resorts spent $425 million modernizing its technology stack over eight years, moving operations to Amazon Web Services. The company achieved a 93% cloud optimization score, above the industry average of 70%-85%.
That infrastructure investment positioned Wyndham to deploy AI efficiently. Connecting data to multiple large language models cost $100,000-a sum CEO Geoffrey Ballotti described as nominal. The company offset this with free AI tools that reduced call center volume and increased upselling.
Five percent of Wyndham's hotels save an average of $61,000 annually through AI-powered tools. One property saved $120,000.
Independent hotels face a different barrier
For smaller operators, the obstacle isn't technology cost-it's preparation. Fifty-eight percent of U.S. hotel leaders worry about sharing data with AI tools, according to research from Access Hospitality.
Many hoteliers understand they need AI but struggle with implementation. Some must build internal knowledge. Others hire consultants or specialists to bridge the gap between hospitality operations and modern technology systems. A new role is emerging: the "hospitality engineer," who understands both guest experience and technical integration. Hiring that expertise costs more.
Sandrine Zechbauer, chief marketing officer at RMS, said many operators lack clarity on how to integrate AI into existing hotel tech stacks. The learning curve and hiring costs are substantial but often invisible in budget planning.
Design decisions determine success or failure
Vittoria argues that a fully agentic approach across all operations is the wrong strategy. Hybrid models-using AI where it delivers clear value and maintaining existing systems where they work-are more sustainable.
The vision for travel's future is an interconnected ecosystem where airlines, hotels, travel management companies, and online travel agencies communicate directly through AI agents. But reaching that vision requires careful design from the start.
Vittoria has spent 18 months making this argument: "We cannot make the mistake of thinking AI is going to solve everything. Then we get a $50 million electricity cost from AWS and shut it down, just because we designed the wrong solution."
For hospitality leaders, the takeaway is straightforward. Evaluate AI based on what it replaces and what it costs to implement, not on vendor promises. Build infrastructure before deploying agents. Budget for talent and preparation, not just software licenses. Learn more about AI for Hospitality & Events and AI for Executives & Strategy to develop a cost-aware implementation approach.
Your membership also unlocks:
