The working reference for the AI-search era β how to get found, extracted, and cited by the engines that now answer questions instead of just linking to them. Primary-sourced, myth-corrected, and read from the papers the field never opened.
Search changed. It stopped being only about ranking a blue link and became about being named inside a generated answer. Most SEO advice hasn't caught up β it repeats blog claims the primary research contradicts. Every figure in this wiki traces to a primary source and is stamped with the era it came from.
What this wiki is
The Bliss SEO Wiki is a working encyclopedia for winning AI search β the reference we use to make our own clients citable, published in the open. It spans the full stack, from whether an AI crawler can physically reach your page to whether an autonomous agent can transact with it. It is deliberately not a list of tips: every claim is traced to a primary study, dated to its era, and corrected the moment newer evidence overturns it.
The difference from everything else written about "AI SEO" is the sourcing. Most of the field cites other blogs, which cite other blogs, until a number nobody measured becomes gospel. We read the actual papers β the arXiv studies, the controlled tests, the official crawler docs β and publish what they really say, including the many places the popular advice is backwards.
The five disciplines
AI search is five layers, not one skill. Each hub below is a self-contained track β start with the one that matches the problem in front of you.
Most "we're not getting cited" problems are access problems. Start in SEO Foundations β if an AI crawler can't fetch and render your page, nothing downstream matters.
2
Make your evidence liftable
Then GEO. Fact density β verifiable statistics, dates, and named entities per 100 words β is the single strongest measured lever for earning a citation inside a generated answer.
3
Build the authority engines trust
Then Off-Page: earned mentions and cross-source consensus are what move an engine from "found you" to "cites you."
How we choose what to cover
Every topic here is scored by the Bliss Authority Score β a model that weighs a topic's real impact, how winnable it is, and how badly the field currently gets it wrong, rather than raw search popularity. That last part matters: when we pulled search-volume data on 100 candidate topics, 74 had effectively none. Nobody is Googling "citation-rewrite retrieval tradeoff" β yet it decides whether you get cited. Popularity is the wrong compass for a frontier reference, so we don't use it as one. The full method is public β see How We Rank.
A living reference. New research overturns old advice constantly. When a primary source corrects something we've published, we correct it in the open β the myths we've already debunked are the point, not an embarrassment.
