AI and Travel: Are We There Yet?

Timothy O’Neil-Dunne discusses why the travel industry shouldn’t be over-reliant on artificial intelligence just yet.

Artificial intelligence is seemingly becoming the solution to everything – but I am skeptical, especially in travel. There is a risk of over reliance on the outcome of AI. My conclusion is that in the world of travel, AI is being overhyped. Below I explore two themes that are specifically aimed at travel.

The first thing we should do is to eliminate the use of the term “artificial” My preferred alternate word is “assisted.” It really does give the right level of meaning and allows for lots of growth and progress.

Theme 1: Big tech investment and its impact on startups

The investment landscape for travel startups has seen a shift, especially post-pandemic. While fewer travel startups are attracting funding, those that do are securing substantial amounts, with average funding per round increasing from $4 million in 2010 to $20 million in 2022.

This trend suggests that while funding is more selective, tech giants and investors are willing to back established players or those demonstrating profitability. The consolidation of the market has made it tougher for new entrants, as investors prefer to fund category leaders over smaller companies with limited disruption potential. I encourage you to check out the PhocusWright Innovation platform.

Tech giants like Amazon, Google and Microsoft have poured billions into cloud infrastructure, data analytics and AI technology. These investments create an ecosystem where startups can access resources that were once reserved for industry behemoths. Madrona Venture Group highlights how large-scale cloud investment empowers startups, offering the kind of computational power and storage necessary to develop cutting-edge travel technologies.

However, McKinsey’s research focus on AI reveals that this investment wave favors mature startups, leaving younger and less-proven ventures on shaky ground. Although hospitality startups remain the leading category for funding, other sectors, such as pre-trip planning and customer service, receive far less investment. Additionally, only 5% of travel startup funding in 2021 came from travel companies themselves, reflecting an over-reliance on tech giants and venture capitalists to drive innovation.

This approach has led to high-profile failures. For example, AI-driven travel management startup Lola, which had all-star team, shut down despite raising more than $80 million in venture funding.

Similarly, Hitlist, which leveraged AI to offer personalized travel deals, faded due to a lack of effective monetization. These cases highlight that investment doesn’t always equate to success. The travel industry’s tendency to prioritize promising technology often overlooks its practical limitations, resulting in a landscape filled with both winners and high-profile failures.

Theme 2: Are digital agents the answer?

The travel industry has always relied on agents, either through direct customer service or third-party travel agents. With the advent of digital agents (AI chatbots, virtual assistants), there’s a promise of a seamless, automated experience. However, the risk again lies in over-reliance. Humans still excel in providing nuanced, empathetic responses and managing complex scenarios – traits that AI struggles to replicate fully. Talk to people trying to make this work. The hard truth is a long way from this vision.

Part of this challenge is the multi-dimensional nature of travel. Humans are good at seeing and connecting patterns that logic might miss. This is an area that fascinates me. I was involved in NORA [non-obvious relationship awareness], which really expanded my own mind. AI models like ChatGPT and Gemini offer powerful tools for automating certain aspects of customer service, but they fall short in situations requiring contextual understanding and emotional intelligence. Large Language Models (LLMs) are mostly generic and this lack the specific travel context. There’s promise in those using LLMs on private data. This will take a lot of upfront learning. 

One challenge to caution everyone on is a two-edged sword. Too much focus brings self-fulfilling prophesies. (Loads of experiments on this topic). The other challenge is that the bots are begetting their own bots. And just like social media we end up in echo chambers.

Digital agents might handle basic queries, but the value of human oversight remains essential. The risk of travel companies blindly leaning into AI is the potential loss of personalized service, which has always been a cornerstone of the travel experience. The industry risks overhyping AI as a catch-all solution without accounting for these limitations. Pattern recognition needs to advance and more power to those working in that area. Check out this piece from McKinsey for a deep dive into the topic. 

Conclusion: AI in travel – overhyped or the future?

Duh… of course AI will have a role in travel. But AI’s role in the travel industry is, and for a very long time will be, complex. While large-scale investment by tech giants can boost startups, it has not necessarily fostered a broad, competitive ecosystem.

Digital agents offer convenience but at the potential cost of the human touch. In travel, where customer experience is paramount, the overhyped nature of AI becomes evident. For AI to truly benefit travel, it must be seen as a tool to enhance human agents, not replace them.

For startups, please stop trying to solve the biggest problems.

• You are already too late, and 
• You don’t have enough breadth and depth to make it. 

Where I do see startups becoming much more relevant is in solving core problems in the middle and the back of the tech stack. One final plea from me … find a way to bring trust to results!

Timothy O’Neil-Dunne is the principal at T2Impact, LLC. Connect with Timothy on LinkedIn.

This article originally appeared on PhocusWire.