Use AI for Itinerary Research Without Losing Care

An agency owner told me recently that she had cut her Portugal itinerary research time from six hours to ninety minutes using AI. Then she said something that stopped me. "But my Portugal itineraries got worse for about a month before they got better." That is the honest version of the AI conversation nobody puts on a webinar slide.

The tools are real. The time savings are real. And the risk of quietly draining the humanity out of your work is also real. This is not a "AI is coming for advisors" post—advisors who steward relationships are not being replaced by chatbots. This is a "how do I use these tools without becoming forgettable" post, because forgettable is the actual competitor.

Where AI genuinely earns a seat at your desk

There are a handful of research tasks where AI is not just fast, it's better than the way most of us used to work. From my seat at WorldVia I watch advisors do wonders with it in four zones.

• First-pass destination scans. "Give me a comparison of three Douro Valley wine regions for a couple in their fifties who prefer intimate over crowded, with day-trip access from Porto." You get a starting map in ninety seconds.
• Supplier and hotel long-lists. Not the final choice. The long-list. Pair AI output with your consortia tools and your own recent site inspections.
• Rewriting technical supplier language into human-readable client language. A dense fare rule becomes a sentence a real person can act on.
• Restaurant and activity clustering by neighborhood. "Here are seven dinner options within walking distance of the Alfama hotel, sorted by noise level." That kind of clustering used to take three tabs and a headache.

The advisors I talk with who use AI-assisted research describe saving a real chunk of prep time, depending on the task. That variation is honest. AI is not a single tool—it's a set of tools that behave very differently depending on what you ask.

Where AI quietly makes your work worse

Here is the part I don't hear said out loud enough. AI is confident. Confidence is not accuracy. In an industry where a wrong airport transfer or a hallucinated room category can cost your client their anniversary, confidence without verification is a liability.

Three specific places to be careful:

1. Anything with a date, a price, or a policy. AI models make things up here with a straight face. Verify with the supplier every time.
2. Anything niche or seasonal. Small boutique hotels, festival timing, ferry schedules, private guides. AI's training data thins out in exactly the corners where your value lives.
3. Anything emotional. The pace of a trip. The right morning to be gentle. The dinner that becomes the story. AI can suggest. Only a human who has listened can decide.

An advisor sent me a screenshot of a Kyoto itinerary a client had built with a chatbot before hiring her. Three restaurants had closed. One had never existed. The chatbot cheerfully placed a fictional teahouse near a real temple. The client's takeaway was, "I need someone who has actually stood in these streets." That's your moat.

A four-step workflow that keeps you in the driver's seat

Here is a workflow I've seen work across advisors who have made peace with the tools. Simple is the point—complexity is how AI sneaks into places it doesn't belong.

Step 1: Human intake first, always. Never open a chatbot before you've had the client conversation. The pace, the sensitivities, the daughter-in-law who is quietly nervous about flying. That is not AI territory.

Step 2: AI drafts the skeleton. Ask it for destinations, neighborhoods, hotel long-lists, and day-flow. Feed it your client notes in your own words. Treat the output as a first draft written by a fast intern who has never met the client.

Step 3: You verify and personalize. Every hotel confirmed live. Every restaurant checked for current hours. Every day rewritten with the client's rhythm in mind. This is where your consortia relationships, site inspection memory, and taste live.

Step 4: You add the sentences only you could write. "I chose this hotel because the concierge is a former sommelier and I remembered you asked about port on our first call." These sentences are the reason clients refer you. AI cannot write them, because AI wasn't on the intake call.

An advisor at our last owner event put it this way: "AI writes the outline. I write the love letter."

What personalization actually looks like on the page

Personalization is often confused with putting the client's name on things. That is customization. Personalization is when a client reads their itinerary and thinks, "she listened to me."

A few tangible ways that shows up:

• A note under a restaurant reservation that says "quiet corner requested, they know your dad uses a cane"
• A morning marked "optional slow start—no plans until 11" because the client mentioned jet lag anxiety
• A specific taxi driver's name in Lisbon because you use the same one every time and he brings a smile
• A sentence at the bottom of Day 3 that reads "this is the day I picked for your anniversary dinner—the sunset from the terrace is worth the walk"

None of that is AI-generated. All of it is what your client is paying for. The AI just cleared desk space for you to write it.

An invitation before you open the next chat window

The advisors I watch thrive with AI are not the ones typing the fastest prompts. They are the ones who protect the parts of the work only they can do. AI is at its best when it hands you back time you can spend on the human parts of your craft. What might change in your next itinerary if you let AI handle only the skeleton, and reserved the last thirty minutes for the sentences only you could write?