What Agentic AI Means for the Future of Trip Planning — And Why Your Travel Agent Might Be Obsolete

Between Expedia’s partnership with OpenAI’s Operator and Google’s covert introduction of Gemini travel agents, the trip-planning industry evolved into something more bizarre than most tourists have realized. There was no big product launch or press conference to accompany the change. A new button on Instagram that converts a Reel into an itinerary and a Copilot feature that discreetly makes hotel reservations while you’re responding to emails are just two examples of how it infiltrated. The industry seems to be changing in real time, and most of us still think of AI as a chatty research assistant instead of what it’s truly evolving into.

Speaking with AI has been like consulting a well-read friend who never quite gets things done for almost three years. It provides you with a five-day itinerary for Lisbon, but you still need to open six tabs in your browser to make the reservation. That pattern is broken by agentic AI. It suggests more than just the trip. It purchases it. When a storm strikes Frankfurt, it reschedules your flight for two in the morning, works out a late check-in with the hotel, and only pings you when something really needs your attention. People consistently undersell that leap.

According to Max Starkov, a hospitality technologist cited in recent CNBC reporting, this change represents the largest shift in online travel since the internet’s inception. Although it’s a strong assertion, it’s difficult to completely discount in light of the rollouts. AI search on Google’s homepage is being tested. Expedia has already integrated OpenAI’s Operator, which was introduced in January. The same is being done by Microsoft Copilot Actions. The infrastructure is being built more quickly than the discourse surrounding it.

Context distinguishes agentic AI from the chatbots we’ve become accustomed to. A conventional booking website presents a wall of choices after requesting dates and a budget. Theoretically, an agent is aware that you have two Marriott stays this quarter, that you have to be home on Wednesday for your daughter’s recital, that you detest red-eye flights, and that your favorite band is performing in Austin on Thursday. This kind of layered understanding is what Jay Richmond at Amadeus called the “missing piece” in personalization, and he’s probably right. You were never truly known to the old model. Maybe the new one.

This leads us to the travel agent. Not the software firm. The real person who has spent twenty years learning which Greek islands are quiet in September, usually from a spare bedroom or a small office above a dry cleaner. For simple travel, such as a corporate reservation to Frankfurt, a long weekend in Barcelona, or a Disney week, the agentic tools already perform the majority of an agent’s duties in less than a minute. In a few years, the typical customer might no longer see the value of contacting a human.

However, there is a more subdued argument that is worth presenting. The limitations of automation are evident to anyone who has attempted to resolve a poorly executed reservation via a chatbot at 11 p.m. The messy, costly, emotionally charged trips—honeymoons, multigenerational safaris, and once-in-a-lifetime itineraries where someone wants a real voice on the phone when things go wrong—are likely to be handled by the travel agents who survive. In a recent LinkedIn post, Olly Nicholls succinctly stated that your value proposition will determine your future rather than agentic AI.

It’s actually unclear if most agents will make the necessary adjustments in time. Technology is advancing more quickly than people can retrain to pursue careers in it. The summer travel season is quickly approaching, and a sizable portion of reservations this year will be made by non-human entities. Most tourists won’t be aware of it. That’s the story, in a sense.