How AI answers pick brands
Someone deciding what to buy increasingly asks an assistant first. They type "best emergency plumber in Austin" into ChatGPT and get back three or four names with a sentence each. If your brand isn't one of them, you never entered the decision. That question never reached your site, never showed up in your analytics, and never appeared in your rankings.
Answer Engine Optimization is the work of being one of those names.
An assistant writes the answer from what it reads#
It draws on two things: what it absorbed during training, and whatever it looks up at the moment you ask. Neither is a ranking.
Three consequences follow.
There is no position one. Being named at all is the first hurdle. Being named early in the list is the second.
The sources are not your site. When an assistant looks something up mid-answer, it reads whatever it trusts on the topic, usually a roundup post, a forum thread, or a review site. Being excellent on your own domain doesn't put you in the answer. Being named on the pages the assistant reads does.
The answer changes without warning. Two people asking the same question an hour apart can get different brands. A competitor publishing one well-aimed page can drop you out. There is no stable position to check once a quarter.
Three things you can change#
In the order they pay off:
- Be present in the sources. Get named on the pages assistants already read for your category.
- Be readable. If an assistant fetches your page and gets an empty JavaScript shell, nothing else you do matters. That's the Pre-rendering chapter.
- Answer the question directly. Pages that answer a specific buyer question in plain terms get quoted. Pages about your company do not.
Before you start#
You need a site added to Encited and a brand book generated for it. The brand book tells Encited which mentions are yours, who your competitors are, and what you sell. Everything in this chapter reads from it, so a wrong brand book produces confidently wrong numbers.
Available on all plans Your plan sets how many prompts you can track and which engines you can pick. Fan-out queries need Pro or above.