Getting your restaurant or hotel website ready for AI search
I've been asked about this more in the last few months than in the previous two years put together, usually in the same way: is AI visibility something we now need to be paying for?
I've just spent a few weeks doing the work properly on my own website, so I thought I'd share what I found. Not theory, and not a sales pitch for a new service I've suddenly invented.
Here's what actually matters, and what I think is being oversold.
First, the thing nobody tells you
AI answers often come from search results.
When someone asks an AI assistant a question, it can run a live search, read the pages it finds and build an answer from them. Which means the foundations haven't suddenly changed. If your website isn't easy to find and understand through normal search, you're already making life difficult for AI search too.
That's worth saying plainly, because there's a lot of noise suggesting this is a completely separate discipline requiring a whole new strategy. In reality, much of it is the same good website and SEO practice, plus a few specific considerations.
What is different
Three things.
Answers are written, not just listed. A traditional search result gives someone a list of links and lets them choose. An AI answer might pick a handful of businesses and describe them. So the question isn't only whether you appear. It's whether a machine can understand your website well enough to write an accurate sentence about you.
Specific facts get repeated. Vague ones don't. "Open from 7am, breakfast until 11" is useful information that can appear in an answer. "Serving the community since 2019 with passion" isn't. The same goes for location, price range, what you're known for and the kind of experience you offer.
Structure carries meaning. A page with proper headings, questions answered directly and clean underlying code is far easier to understand than a page of beautifully written prose with no obvious signposts.
What to actually do
Six things, roughly in order of how much they matter.
Say what you are, plainly, near the top. Somewhere in the first paragraph of your homepage, state what the business is, where it is and what you offer. "A neighbourhood Italian restaurant in Hackney, open for dinner Tuesday to Sunday." It feels obvious to the point of being dull, but that's partly the point. A machine can't infer atmosphere from a metaphor.
That doesn't mean stripping all the personality out of your brand copy. You can still have the beautiful headline and distinctive tone of voice. Just make sure the useful information is there too.
Add schema markup. This is structured information within your website code that can help search engines understand things like your business type, address, opening hours, price range and menu. For hospitality businesses, there are specific schema types built for exactly this.
It's invisible to customers, but it's a useful part of making your website easier for search engines and other systems to understand.
Answer real questions on the page. Do you take walk-ins? Is there parking? Can you host a private dinner for twelve? Do you cater for children? These are exactly the kinds of questions people ask AI assistants, so it makes sense for your website to answer them clearly.
An FAQ section is a natural home for this, and hospitality websites tend to need one anyway. Write each question as a customer might actually ask it, with a simple answer underneath. It helps people and makes the information easier for machines to understand too.
Get the basics right. Consistent name, address and phone number everywhere. An up-to-date Google Business Profile. Useful alt text on images. Fast loading. None of this is new advice, but all of it still counts, and most of it should be handled properly when a site is built well in the first place.
Make sure you exist elsewhere. This is the part people often skip, and it's incredibly important. Search engines and AI systems don't have to rely solely on what you say about yourself.
Local press, review platforms, restaurant guides, directories, awards and other relevant websites all help build a clearer picture of who you are. If the only place talking about your business is your own website, you're relying heavily on one source.
Then check. Ask an AI assistant the kind of question your ideal customer might ask. "Where should I eat in Hackney tonight?" "What are the best boutique hotels in Bath?" See who comes up and, where possible, look at the sources being referenced.
That can tell you a lot about where the gaps are.
What I found doing this on my own site
I ran ten different questions across two AI assistants.
When I asked about me directly, both described the business accurately, which was reassuring and suggested the work I'd done to make the site clearer was helping.
When I asked the kinds of questions a potential client might actually type, it was patchier. I appeared for some and not others.
And that was probably the most useful finding.
The gaps weren't necessarily something I could fix with another line of code. Other businesses had more presence elsewhere. More mentions, more third-party sources and more places online reinforcing who they were and what they did.
The work on my own site helped AI understand and describe my business. Being recommended is a bigger picture.
What I'd ignore
Anyone telling you AI optimisation is an entirely separate world. A lot of what's currently being discussed as AI optimisation overlaps with good SEO, good website development and clear content. There are new considerations, but I wouldn't throw everything you know about search out of the window.
Writing for machines instead of people. The old version of this mistake was keyword stuffing. The new version is stripping all the character out of your copy so it reads like a database entry.
Your website still has to convince an actual person to book a table, reserve a room or make an enquiry. Being understood by a machine is useful. Being compelling to a human is still the point.
Chasing every new tool. This space is moving incredibly quickly. I wouldn't build your whole digital strategy around whatever new acronym or optimisation tool has appeared this month.
The durable things are clarity, structure, authority and being known somewhere other than your own website. They mattered before AI search and I suspect they'll matter for a long time afterwards.
The honest summary
Most of what makes a website work well in AI search is what made a good website work before: say clearly what you are, answer the questions people actually ask, get the technical foundations right and build a reputation beyond your own website.
What's changed is that vagueness can cost you.
If an AI assistant is trying to write one useful sentence about your restaurant or hotel, it can only work with what it can find and understand. If the website looks beautiful but says almost nothing specific, you're making that job much harder.
And if the real issue is that the business itself hasn't decided what it stands for, who it's for or why someone should choose it, no amount of technical work is going to solve that. That's a positioning problem and it's worth getting right first.
If the brand is clear and the website is letting it down, that's a much more straightforward job.
Tell me about your website
If you're opening, replatforming, or working around a website that isn't doing enough, tell me where you are and I'll come back to you within a working day.
