Former travel writers test AI trip planning in France

By AI, Created 6:06 AM UTC, June 05, 2026, /AGP/ – The team behind Food & Drink Destinations used Gemini and Perplexity to plan a trip across France and found the tools about 80% accurate. The experiment suggests AI can handle trip structure, but travelers still need human judgment to verify details.

Why it matters: - The experiment shows where AI is already useful for travel planning and where it still breaks down. - For travelers, that means AI can help shape an itinerary, but verification still matters for restaurant hours, menu details, and niche experiences. - For travel publishers, the result points to a split future: AI may take over logistics and list-style content, while first-hand reporting becomes more valuable.

What happened: - Amber and Eric Hoffman of Food & Drink Destinations asked Gemini and Perplexity to plan a full trip across France. - The Hoffmans ran the tools in parallel and used roughly 100 queries about their travel history, preferences, and constraints. - Both tools independently recommended the same route: Bordeaux, La Rochelle, and Nantes. - The experiment followed the couple’s decision to leave travel writing around 2022 and 2023 as AI-generated content began flooding search results.

The details: - The Hoffmans spent more than a decade as food and travel writers and worked with tourism boards across Europe. - Their review found AI was about 80% accurate overall. - The tools performed best on the big picture, including choosing a country they did not already know well, sequencing three cities, and mapping a workable train route. - AI also surfaced destinations the Hoffmans say they would not have found on their own, including La Rochelle. - The couple says La Rochelle was unfamiliar even to well-traveled friends and French chefs they know. - Restaurant recommendations were right roughly 80% of the time. - AI pointed them to a Bordeaux food market they returned to three days in a row. - The tools struggled with granular information, especially restaurant hours and menu details. - AI could not reliably reproduce the eat-in market experience they found in Bordeaux when asked for similar options in other cities. - In one case, the tools suggested a wine tasting that did not exist. - The full account covers Bordeaux, La Rochelle, Nantes, and a day trip to Clisson. - Food & Drink Destinations publishes first-hand culinary guides across Asia and Europe. - Amber S. Hoffman also writes about AI in business and consults with companies on AI visibility. - The experiment tested not only trip planning, but also how well AI surfaces and recommends real-world hospitality businesses. - Food & Drink Destinations published the full account.

Between the lines: - The Hoffmans are not arguing that AI is useless. - Their findings suggest AI is strongest as a research and routing tool, not as a substitute for on-the-ground judgment. - The project also reflects a broader concern among writers who have watched AI reshape search, content discovery, and travel inspiration.

What’s next: - Travelers are likely to keep using AI for itinerary building, then fact-checking the results before they go. - Travel publishers may lean harder into reporting, personal experience, and local discovery as differentiators. - The Hoffmans plan to keep examining what AI can and cannot do for real-world travel decisions.

The bottom line: - AI can build a good trip framework, but human expertise still wins on the details.