Insights · Field note 01
It isn't a ranking, and it isn't random. Assistants assemble an answer from what they were trained on and what they can retrieve right now — and only a handful of names survive.
When someone asks an assistant to recommend a business, the model doesn't scroll a list of links. It writes a short paragraph from two sources: the patterns it absorbed in training, and whatever it can fetch live from the open web at that moment. Your job is to be legible to both.
That legibility breaks into four honest signals. Can the model read your site at all? Does it have reason to trust you exist? Does it actually name you when asked? And when it does, is the description accurate and flattering? Most businesses quietly fail the first two without ever knowing.
There is no registry to submit to, no button that puts you in the answer. There is only the unglamorous, decisive work of making the public web state your facts clearly and consistently — which is precisely what a AI HALO engagement does.
Continue reading