Why ChatGPT Can Plan A Trip, But It Can’t Replace Your Travel Counsellor!
A “perfect” AI itinerary still needed a human touch to get it right
Recently, a family member came to me very excited about a multi-centre trip to Finland next year. He and his partner are in their 70s and wanted a week away to see more than “just Helsinki”: lakes, proper Finnish culture, gentle exploring, and crucially, avoiding the midges and mosquitoes. He proudly told me he had “done all the research with ChatGPT”, had a “brilliant itinerary” and was about to book it all himself. I smiled and said, “Why don’t you let me take a look? I’m happy to review it and just sense-check how it fits with what I know you’ll actually need.” Once we dug into what they really wanted, a few cracks in the AI‑designed plan appeared quite quickly.
1. Seasonality and what’s actually open
On paper, the suggested lake-region stay looked wonderful: peaceful scenery, classic Finnish landscapes, lovely hotel. The problem was timing.
They wanted to avoid peak mosquito season but still enjoy activities and local experiences around the lakes. The itinerary sent them there outside the main operating period. In reality, that meant:
- Limited activities and excursions
- Lots of things closed or on reduced hours
- A much quieter atmosphere than they had in mind
So yes, “Finland lakes” was technically ticked, but the experience would have been underwhelming. This is where human knowledge counts: knowing when somewhere is open, and when it is actually worth going.
2. Trains, transfers and what is realistic in your 70s
They were very clear that they did not want to hire a car. They wanted to travel by train and keep things simple and comfortable.
AI happily plotted out multiple stops that looked perfectly joined up on a screen, but:
- Some destinations were awkward to reach by rail
- Others involved complicated changes and transfers
- It just was not realistic or enjoyable for two people in their 70s with luggage in tow
Could you technically get everywhere? Yes.
Would it feel like a faff and a half? Also yes.
I reworked it using Finland’s rail network in a much more direct, sensible way, choosing stops that made sense both on a map and for how they like to travel.
3. Pace and energy levels
Seven nights is not a huge amount of time once you factor in travel, settling in and actually enjoying yourself.
AI had done what it does best: maximised “stuff”. Lots of places, lots of moves, lots of box‑ticking. Impressive on screen, exhausting in real life.
I slowed everything down. Looking at:
- The best times of year for the regions they liked
- Distances and timings by rail
- Their age, energy levels and desire for a “gentle” adventure
I created a softer, more thoughtful pace that still gave them variety in culture, history and scenery, without needing a lie-down by day three.
4. Curated hotels, not random listings
Then there is the hotel piece.
AI can suggest hotels but it cannot:
- Draw on independent reviews from trusted colleagues
- Balance location, comfort and character for this specific couple
- Spot the little red flags that only come from real client feedback
I gave them a small, well‑researched shortlist of places to stay, all chosen with their tastes and needs in mind. Not just “4.5 stars online”, but “this actually feels right for you”.
5. One joined-up booking, with protection and support
Finally, the fun bit: what happens when things change.
The AI version had them booking each bit separately: trains, hotels, extras, all DIY. If anything moved or went wrong, they would be left doing travel‑Tetris on their own.
I was able to:
- Package the trip properly, with protection and peace of mind
- Take care of all the rail bookings and seats
- Make sure timings worked comfortably as a whole
- Be their one point of contact if anything needed to change
They did not just get an itinerary. They got a professionally designed, fully supported holiday.
So, is AI going to steal my job?
Not any time soon!
My job is to:
- Listen to what clients really want (and what they definitely do not want)
- Turn vague ideas into realistic, well‑paced, personally tailored trips
- Know when somewhere looks great online but will, in reality, be shut, quiet or awkward to reach
- Match hotels and experiences to real people, not generic “users”
- Pull flights, trains, transfers and stays into one coherent, protected journey
- Be there before, during and after the trip when questions or wobbles appear
I use AI daily. It is brilliant for sparking ideas and speeding up parts of the process. But it does not know my clients: their health, their quirks, their “absolutely nots”, or the little touches that turn a nice break into “best holiday we have ever had”.
In the Finland example, AI created a starting point. My expertise turned it into something that would actually work for two real humans in their 70s. They were delighted and booked straight away.
So no, AI is not here to steal my job. It is here to sit in the background, while I ask better questions, challenge its suggestions and design trips that fit real people, not just clever algorithms.
