Paul Bruemmer

05 — Questions

Questions

Including questions on GEO and AI search drawn from his ongoing research and writing.

Is GEO a new discipline, or SEO rebranded?

Both, but not equally. A lot of what people now call GEO is still the work good SEO should have been doing all along: clear technical access, structured and consistent information, credible sourcing, entity clarity, and content that directly answers real questions. What has changed is the output environment. In traditional search, visibility often meant earning a ranked link. In AI search and answer engines, visibility increasingly means being one of the sources an assistant understands, trusts, summarizes, or cites when it generates an answer. So GEO is not a clean replacement for SEO. It is closer to an extension of SEO into answer-based discovery.

How do AI assistants decide which sources to cite?

Not fully knowable from the outside. The exact selection logic is not public, and it changes by system, query, model, index, and retrieval method. But through testing, we can observe patterns. AI assistants tend to do better with sources that are clear about who they are, what they do, where they operate, and why they are credible. They also tend to favor information that is corroborated across independent sources, technically accessible, easy to extract, and written in passages that can be cleanly summarized or attributed. That does not mean citations can be guaranteed. It means the source can be made easier to understand, verify, and use.

What can a company actually influence, and what can't it?

A company can influence how clearly it describes itself, how consistent its information is across the web, whether its important pages are crawlable, whether its facts are supported by visible evidence, and whether its content is structured in a way that machines and humans can both understand. A company generally cannot control a model's historical training data, private weighting systems, competitor comparisons, or whether an assistant chooses to cite anyone for a particular answer. That is why any promise of guaranteed AI citations should be treated skeptically. The real opportunity is not control. It is improving the quality, consistency, accessibility, and credibility of the evidence AI systems can find.

How long has Paul Bruemmer worked in search?

Since 1994. He founded WebSpecialist LLC (WebSpecialist.com) in December 1994 and served as its managing partner through May 1997. His subsequent experience spans paid search, the algorithmic era, enterprise in-house programs, and now AI search and generative engine optimization.

What is Paul Bruemmer known for?

Building the enterprise SEO function at Overstock.com from zero between 2008 and 2011, and leading the company through a precedent-setting Google penalty and recovery. Approximately 200 published articles for ClickZ beginning in 2000, alongside Danny Sullivan and Bruce Clay, with later bylines in Search Engine Land and MarketingProfs. And early writing on search becoming an answer engine, published in Search Engine Land in April 2012.

What does he work on now?

Primarily research, writing and SCORE mentoring — on how search is evolving, including AI search, where practical AI is useful, and what the evidence supports. He also follows quantum computing and its potential convergence with AI as a research interest. Alongside that he builds and improves tools and experiments, including AI Search Checkup, which assesses how AI and search platforms interpret an organization and converts the findings into prioritized remediation; Entity Level Authority, an entity-authority and AI-discoverability framework for plaintiff law firms; Fact Check Weekly, an AI-assisted claim review and editorial workflow for newsrooms; and CatchDesk.ai, bilingual AI call triage and lead capture for service contractors.

Does he build his own tools, or only advise?

Both. Alongside research, mentoring and occasional advisory work, he builds his own diagnostics. The most developed is the Semantic Graph Optimizer, an internal crawler-perspective analyzer that separates what a website establishes on its own — the entities and relationships a machine can find in the visible evidence — from the associations a business has confirmed it wants to be known for, and reports where the two diverge. It is an internal tool used in advisory work, not a product, and is not offered for sale or public access. A lighter public companion, Findable, offers a preliminary, plain-English check for anyone getting oriented. Neither promises rankings or citations; both measure the same thing — whether a site's crawler-visible information gives search and AI systems a clearer foundation to interpret the organization.

Is Paul Bruemmer available for advisory work?

Selectively. When availability and fit allow, he takes on clearly defined SEO and AI advisory projects — for example a search strategy review, an assessment of a proposed AI application, or a prioritized set of recommendations — each with an agreed scope, deliverables and endpoint. SCORE mentoring is separate from paid advisory work.