llms.txt, schema and NAP consistency: what actually gets you cited in 2026

There’s a tidy list doing the rounds of what an independent property “needs” to be visible in AI search: add schema, fix your NAP, publish an llms.txt file.

It’s repeated so confidently that owners assume all three carry equal weight. They don’t. Two of them are quietly essential and proven. One is a reasonable bet on the future that, as of 2026, the major AI systems mostly ignore. It’s worth knowing which is which before you spend money, so here’s the honest version.

NAP consistency: dull, and the one that matters most

NAP stands for Name, Address, Phone: the basic identifying facts of your property. NAP consistency means those facts appear identically everywhere you exist online: your website, your Google Business Profile, every directory, every OTA listing, every review site.

It sounds trivial. It is anything but, because inconsistency actively harms you. When your address is formatted three different ways across five platforms, or an old phone number lingers on a directory you forgot about, both search engines and AI assistants lose confidence that these listings describe the same, real, single business.

A system that isn’t confident about who you are will not confidently recommend you.

It reaches instead for the property whose details line up cleanly everywhere. This is the least glamorous work in all of digital, and for a small property it’s frequently the single highest-return fix available, not because it’s clever, but because so few owners have ever audited it, and the errors accumulate silently over years.

Schema: giving machines the facts in a language they read

Schema markup is a small layer of structured code added to your website that states your facts in a format machines read directly: this is a hotel, here is its location, here is its price range, here are its rooms, here is its rating.

To a human visitor, schema is invisible. Your site looks exactly the same. To a search engine or an AI crawler, it’s the difference between having to guess what your page means and being told plainly. It’s the tidy, machine-readable version of everything a person would otherwise have to infer from your prose.

This is proven, current infrastructure. It’s how you earn rich results in Google, and it’s a genuine signal that AI systems use when they build an understanding of your property. For a distinctive independent place, schema is where you make your specifics (the chapel, the sea view, the exact location) impossible to miss or misread. It sits at the centre of the wider technical foundations work, and it’s worth doing properly.

llms.txt: the honest truth the hype leaves out

Now the one that needs straightening out. llms.txt is a proposed file you place at the root of your site giving AI tools a clean, curated map of your most important pages. The idea is genuinely sensible, and you’ll see it sold hard as an AI-search essential.

Here is the 2026 reality, stated plainly because you deserve it: the major AI crawlers largely do not use it yet. Independent studies through the year put adoption at roughly one site in ten or fewer, and, more to the point, monitoring of actual AI crawler traffic shows the search-and-answer bots overwhelmingly skip the file and read your normal HTML instead. Google has said on the record that it doesn’t use llms.txt, likening it to a long-discredited old tactic. No major AI provider has committed to acting on it in production.

So should you publish one? On balance, yes, but for the right reason and at the right cost. It’s cheap, it’s harmless done correctly, and if the standard is adopted later you’re already positioned. Think of it as a low-cost bet on where the agentic web may go, not as something that earns you citations today. What it must never do is jump ahead of the work that genuinely moves the needle.

The order that actually works

If you take one thing from this: the boring fundamentals still beat the shiny new file. In order:

The reason to be straight with you about this is simple. The whole point of getting cited by AI is being a property that systems can understand and trust. Publishing content that overstates what a file does, to you or to your guests, is the opposite of that. This piece pairs with our look at how travellers now discover independent hotels through AI assistants; the same foundations decide whether you’re in that answer at all.

The unglamorous work is the work. It compounds, it lasts, and it connects directly to the commercial side: being legible to machines is what turns platform visibility into direct bookings rather than another disconnected channel.

Reserved Hospitality builds and maintains the digital foundations behind independent boutique properties (schema, listings consistency, the genuinely useful technical groundwork) without the hype and without the jargon.