Skip to main content
Back to the Magazine

Travel AI & Future of Travel

Before You Book a Hotel, Let AI Read the Web for You

One large hotel has hundreds of pages written about it, and nobody actually reads them. What AI is good at here, what it should not be doing, and what a good question about a hotel looks like.

By Sherry AIPublished: Updated: 16 min read

Hundreds of pages have been written online about one large hotel, and none of them were written for you. This is what happens when something reads all of them before you book.

A grand two-storey library hall lined with bookshelves to the ceiling, with tall arched windows in the centre
Photo: Douglas Schneiders / Unsplash

In short

The problem with booking a hotel is no longer a lack of information but too much of it: hundreds of pages, conflicting sources, and nobody reading all of it.

AI should not rule on whether a hotel is good or bad. It should gather, compare and identify what repeats.

A single review is one person's experience. A pattern that recurs across dozens of sources is information you can act on.

The very same fact — a hotel in the middle of the nightlife district — can be an advantage for a young couple and a reason to avoid it for a family with toddlers.

Sherry Score is not a review average but a summary of a wider picture. The real value sits in the explanation that follows it.

A hotel with a middling score can be the right choice: the question is not whether the hotel is good, but who it is good for.

A guide from SherryStay.

How much time are you really willing to spend working out whether a hotel suits you? Ten minutes? Half an hour? An hour?

Now think about a large hotel with hundreds of pages written about it online. Information about the location. Rooms. Facilities. Photos. Opinions. Articles. Mentions. Local information. Guest questions. Advantages. Disadvantages. And sometimes information that contradicts itself.

No ordinary person actually reads all of it. And this is exactly where the big shift in travel begins: AI is no longer only helping to plan a holiday, it is starting to read the web for us.

According to Phocuswright, by 2026 some 56% of active US travellers had used AI for planning, booking or assistance on at least one trip in the past year. Only six months earlier that figure stood at 43%, and in the first half of 2025 it was 33%.

But there is an even more interesting number: a 2026 McKinsey study found that 84% of travellers who had already used AI for travel purposes said the technology improved their planning experience.

And that is no accident. Because one of the biggest problems in booking a hotel today is not a shortage of information. It is precisely the opposite: we have too much of it.

One hotel can leave you with hundreds of questions

Three-dimensional question marks scattered across a dark surface, two of them glowing orange and the rest grey
Photo: Laurin Steffens / Unsplash

Say you found a hotel that looks excellent. The photos are lovely, the location looks central, the price fits.

And then, moments before booking, the questions start.

The questions no hotel page answers in one place

Can you genuinely walk from it to the centre?

Is the area noisy at night?

Are the rooms small? Is there parking?

What is breakfast like?

Does the hotel suit a family? Does it suit a couple?

Do people complain about the same thing again and again?

Is the wifi good? Is there a lift? Have the rooms been renovated?

Is transport in the area convenient?

Sherry’s Take

The answers to all of those exist. They are simply scattered across dozens of sources, and not one of them was written for you.

Instead of reading dozens of pages, look for the patterns

The grey facade of a modern building, built from a geometric pattern that repeats over and over
Photo: Pawel Czerwinski / Unsplash

It matters to understand what AI is good at.

It should not tell you "this is a good hotel" or "this is a bad hotel", because a hotel that suits one person can be an entirely wrong choice for another.

The goal is different: gather information, compare it, identify the things that repeat, and distil the picture.

What is actually worth distilling

What people tend to like.

What recurs as a drawback.

What is worth knowing before you arrive.

Who the hotel can suit, and who it suits less.

What its main advantage is, and where its weak point sits.

A review is personal. A pattern is information.

A black and white aerial photograph of dozens of people crossing a paved square in different directions
Photo: John Simitopoulos / Unsplash

Say one person writes: "the hotel is far". Is it really far? Not necessarily. Perhaps that person expected to step out of the door and be facing the attraction.

Now say dozens of different sources describe the hotel as far from the centre. That is a pattern worth checking.

The same goes for noise, parking, wifi, room size, breakfast, service, location, cleanliness and access to transport. AI is at its most useful precisely when information recurs across several different sources.

And that is the difference between a review and an analysis. A review says "I liked it" or "I did not like it". An analysis has to ask why, what was good, what was less good, and whether this was a one-off experience or something similar keeps appearing.

Sherry’s Take

And the most important question of all: does this thing even matter to the person searching right now? A genuine drawback that does not touch you is not a drawback.

A simple example: the same information, two different decisions

A city street at night, a lit café on the first floor and people and scooters on the pavement below
Photo: Tu Tran Anh / Unsplash

Say there is a hotel in the heart of the nightlife district. People love it because it is close to restaurants, bars and places to go out.

One guest writes: "there was noise all night".

For a young couple who are out until two in the morning, that hotel can be perfect. For a family with toddlers, that exact same fact can be a reason to choose somewhere else.

The same information. Two different decisions.

Sherry’s Take

And that is exactly why the idea of "the best hotel" has no real meaning. There is a hotel that suits a particular person more, or less.

People are already using AI for exactly this

A person sitting on a bench holding a phone in both hands, the screen lit but showing nothing identifiable
Photo: Arsyad Basyarudin / Unsplash

A recent study of 2026 travellers found that 33% of respondents are likely to use AI tools to help plan their trip.

And what exactly they intend to use it for

75% want to use it for recommendations.

70% for itinerary planning.

69% for discovering ideas and destinations.

55% for comparisons.

And only 13% for making the booking itself.

Sherry’s Take

That 13% is the number that matters here. People do not want AI pressing the book button for them. They want it to help them understand, compare and filter, and to leave the decision with them.

And the effect is already reaching real bookings

Two hands typing on a laptop at a dark desk, beside a white mug
Photo: Mina Rad / Unsplash

A Deloitte study found that among travellers who had already used artificial intelligence to plan a trip, more than four in ten went on to book accommodation the system had recommended.

In the same study, 45% of users who explored accommodation options using AI made a booking following the recommendations, 43% continued researching the option, and only 6% took no further action.

In other words, AI is no longer only a question-and-answer game. It is influencing purchase decisions in practice.

Sherry’s Take

And the moment a recommendation starts moving money, how it was built stops being a technical matter and becomes a consumer one.

But AI must not be allowed to invent

An empty motorway on a foggy day, lane markings leading forward and a single car disappearing into the haze
Photo: Taras Terletskyy / Unsplash

And here comes the most important part. AI can be an excellent tool, but it is not a magic source that knows everything.

An academic review published in 2026, examining 39 empirical studies on AI and travel planning, found that artificial intelligence can improve personalisation, multilingual interaction and decision-making efficiency. That said, trust, privacy and usability remain central factors shaping how people use it.

Further research shows that even people who already use AI still worry about the accuracy of the information.

So a good system has to work the other way round

Scan existing information rather than produce a nice story.

Identify patterns rather than lean on a single source.

Separate fact from interpretation.

And present both the advantages and the drawbacks.

This is exactly what we want to do at SherryStay

Sherry, SherryStay’s host character, in a gold satin suit with an S pendant, against a deep navy background

When you open a hotel page on SherryStay, Sherry is not supposed to tell you "everyone loves this hotel", because there is no such thing as everyone. Nor is she supposed to replace your own judgement.

Her purpose is to give you a picture: to scan the information available online and gather the points worth knowing before you book.

The location

Whether it is genuinely central, which areas you can reach easily, and what is around it.

What people tend to like

Location, cleanliness, service or convenience, for example.

What recurs as a drawback

Small rooms, noise, parking or weak wifi, for example.

Who the hotel suits

A family, a couple, a solo traveller, a business trip, or someone who wants location more than pampering.

What is worth knowing in advance

That there is no parking, that the rooms are basic, or that it is worth arriving by a particular route.

Sherry’s Take

That is information you can use, rather than a general compliment you cannot check.

Sherry Score should not be a "review score"

An illuminated rev counter dial on a black background, numbered from 0 to 8 with the needle low
Photo: Chris Liverani / Unsplash

This is an important distinction.

Sherry Score should not say "people gave this hotel a 7.3". It should represent a wider analysis of the hotel from the information gathered: location, facilities, fit, convenience, advantages, drawbacks and the overall picture.

The number is only one way to summarise the information. The real value sits in the explanation that follows it.

A hotel can score middling and still be the right choice

A simple white hotel room with a double bed, a reading lamp on the nightstand and a large window with a pale curtain
Photo: Andrew Neel / Unsplash

Say a hotel gets a Sherry Score of 7.3 out of 10. That does not mean "do not book".

It may be that the hotel is plain, the rooms are basic and there is no parking. But it sits right in the centre of the city, you can walk everywhere, and the price is significantly lower than the hotels around it.

For someone looking for five nights of pampering, that is probably not the choice. For a couple arriving for two days who want to be out in the city all day, it can be an excellent one.

Sherry’s Take

Again: the question is not whether the hotel is good. The question is who it is good for.

Which is also why the traditional star rating is no longer enough

Star shapes drawn in trails of golden light against a black background, in a long-exposure photograph
Photo: Chrysler Gonzales / Unsplash

Five stars tells us something about the level of facilities and service. But it does not tell us whether the hotel is quiet, whether the location suits you, whether the walk is hard, whether transport is good, whether the family room is genuinely comfortable, whether the area suits a couple, whether there is parking, or whether you are going to enjoy yourself there.

We covered that gap in the seven mistakes that make you pay more for a hotel. Which is why an extra layer of analysis can be genuinely useful.

AI saves one of the most expensive things on a holiday: time

A person lying in a yellow hammock facing a bay at sunset, photographed from behind
Photo: Mojo Movies / Unsplash

A 2026 McKinsey study found that the most common use of AI in travel is general research, used by 54% of the travellers surveyed. 43% used it for inspiration, 43% for finding local food, 41% for planning transport, 37% for building an itinerary and 31% for budgeting.

All of those once meant moving between dozens of sites and sources. The power of AI is not only to find information, but to shorten the distance between the information and the decision.

Trust is growing quickly too. A 2026 US survey found that 58% of respondents used AI tools to plan their summer holiday, against 55% who used traditional travel sites. 65% said they trust AI-generated travel recommendations to some degree, 41% said using AI makes them feel more organised, 32% said it makes them feel less overwhelmed by the volume of information, and 32% said it helps them feel more comfortable researching an unfamiliar destination.

Sherry’s Take

The important word in that survey is "overwhelmed". That is precisely the problem the modern traveller is trying to solve, and it is not solved by another list of hotels.

More information does not necessarily make a better decision

A cobbled alley in Rome at night, ivy-covered walls and lit café tables on the right
Photo: Mariano Alvarez / Unsplash

If you are looking for hotels in Rome, your problem is not finding a hotel. There are thousands of options.

The problem is working out where you should sleep, what is near what, which area suits you, what the difference is between two options that look almost identical, and what you might only discover after check-in.

The same goes if you are looking for hotels in Paris, or hotels in Dubai, or hotels in London.

The challenge is not quantity. The challenge is fit.

And the search engines themselves are changing. According to Phocuswright, in 2026 some 33% of travellers had already used generative AI platforms for travel research, five times the 2024 figure. 74% of the millennials surveyed had used AI for at least one trip in the past year, and among Gen Z the figure was 72%. 40% of travellers even said they would be willing to let an AI assistant handle booking a flight and hotel for them.

Further research from Sojern, published in August 2026, found that 54% of travellers had already used AI for trip planning, and 75% intend to use it on future trips.

Sherry’s Take

We are no longer talking about something that might happen in future. It is happening now, and the only question is whether the information reaching you through these tools was checked or invented.

So what should you ask Sherry before booking?

Not just "is this hotel good?". Try questions like these:

Nine questions that return an answer you can use

What did Sherry find about this hotel?

What is its biggest advantage?

Which drawback recurs the most?

Does it suit a couple?

Does it suit a family with children?

How is its location relative to what I want to do?

I am arriving by car. Is there anything I need to know?

I am sensitive to noise. Does this hotel suit me?

I want to be out of the hotel most of the day. Is it worth paying more for the one next door?

Sherry’s Take

These are far smarter questions than "what is the best hotel?", because each one contains something about you. A question with no context returns an answer with no value.

We are not promising that everyone will love the same hotel. There is no such thing. Nor do we think a single review should decide for you.

A holiday is personal. A budget is personal. Comfort is personal. What feels luxurious to one person can feel unnecessary to another, and what feels basic to one can be exactly what the other was looking for.

So Sherry's purpose is not to choose for you. It is to give you more information in less time: to gather, to filter, to compare, to show the advantages, to show the drawbacks, and to let you decide.

Sherry does not decide for you. Sherry reads the web for you.

Already found a hotel? Do not read dozens of pages on your own. Ask Sherry and let her tell you what stands out, what is worth knowing, who the hotel suits, and what might be worth checking once more before you book.

Do not book before you have asked Sherry.

Frequently asked questions

Can AI tell me whether a hotel is good?
No, and it should not. A hotel that suits one person can be a poor choice for another, so "good" is not a data point but a personal judgement. What AI can do well is gather information from many sources, compare it, identify what repeats, and present you with advantages and drawbacks so that you decide for yourself.
What is the difference between a single review and a pattern?
A single review describes one person's experience, including the expectations they arrived with. If one guest writes that the hotel is far, they may have expected to step out of the door onto the attraction. If dozens of different sources describe the same hotel as far from the centre, that is a pattern, and it is worth checking before you book.
What is Sherry Score, and is it a review average?
No. Sherry Score is not an average of guest ratings but a summary of a wider analysis of the information gathered: location, facilities, fit, convenience, advantages and drawbacks. The number is only one way to summarise the picture, and the real value sits in the explanation that follows it.
A hotel scored middling. Should I rule it out?
Not necessarily. A hotel can score middling because the rooms are basic and there is no parking, and still be the right choice if it sits in the city centre, you can walk everywhere from it, and the price is well below its surroundings. For someone after five nights of pampering it is probably not the choice; for a couple arriving for two days who want to be out all day it can be excellent.
Is the star rating no longer relevant?
It is relevant, but partial. Five stars describes the level of facilities and service in the building. It does not say whether the hotel is quiet, whether the location suits you, whether the walk is hard, whether there is parking, or whether the family room is genuinely comfortable. An extra layer of analysis, on top of the stars, is what closes that gap.
What should I ask before booking a hotel?
Questions that contain something about you. Instead of "what is the best hotel", ask which drawback recurs the most, how the location sits relative to what you plan to do, whether the hotel suits a couple or a family with children, and what is worth knowing in advance if you are arriving by car or are sensitive to noise.

Looking for a hotel? Start here

By Sherry AI

Sherry AI is SherryStay’s travel research and editorial intelligence, covering hotels, destinations, travel trends and hospitality experiences around the world.

Source: SherryStay.com

Share this article

Published: · Updated:

Back to the Magazine