The guest list was written years ago.

Somewhere on most business websites is a short text file that decides which machines are allowed to read the site. Most owners have never opened it. Plenty of them were written years ago, by someone who has since moved on, for a web that worked differently.

The file is a guest list. For a long time its job was simple: let the search engines in, keep the scrapers out. Then a new kind of visitor arrived. People now ask an AI assistant which accountant near them is good with small companies, or which agency handles a particular kind of job, and the assistant goes and reads the candidates' sites to answer. That visit happens in the moment, on a person's behalf, about a decision they are making that day.

An assistant reading your page because someone asked about you is a customer arriving through a new door.

A guest list written for the old web often turns that visitor away. Rarely on purpose. A blanket rule against unfamiliar bots. A firewall or bot-protection setting that treats anything automated as hostile. A line copied from a template in another era. The result is a business that is hard to find in exactly the conversations its customers have started having.

Nobody notices, because nothing breaks. The site loads. The contact form works. Search traffic looks the way it always did. The only symptom is an absence: a buyer asked an assistant for three names in your town, and yours was not among them, and there is no report anywhere that tells you it happened. Problems that announce themselves get fixed. This one is silent by construction, which is why so many of those old files are still in place.

A guest list written for the old web turns away the visitors who now answer your customers.

The fix is less technical than it sounds. It is mostly a decision about intent, written down. Sites hosted on our platform say it in so many words: they name the AI search tools, and the assistants that fetch a page when a person asks about it, and they tell them they are welcome. Nothing is disallowed. The crawlers that collect pages to train models are not named at all. They stay covered by the same general welcome the file has always given everyone else.

That distinction is worth keeping. A visitor reading your page to answer a live question from a buyer is doing something different from a crawler gathering text in bulk. The first is a sales conversation you are not in the room for. Naming it, and saying yes to it, is the least a business can do for a conversation about itself.

In the spirit of saying exactly what is true: our own corporate site is plainer. It welcomes every reader through the general rule and names nobody. Nothing is turned away, but nothing is named either.

The difference between those two is small and real. Silence lets a careful visitor in. Naming it removes the doubt.

If your website lives somewhere else, open yourdomain.com/robots.txt and look for rules that turn AI assistants away. Then ask whoever runs your firewall or bot protection the same question, because a setting there can refuse a visitor the file would welcome.

Being readable does not make you recommended. Being unreadable guarantees you are not.

Being readable is not the same as being recommended. What an assistant says about a business depends on what it finds there, and on everything else it finds elsewhere. But it cannot cite a page it was not allowed to read, and it cannot recommend a business it could not see.

Every website has a guest list. Read yours. The newest guests are asking on your customers' behalf.