Skip to content
Dawn Anderson

Dawn Anderson

Founder, Bertey

Bertey

At a Glance

Role
Founder and SEO consultant, Bertey, Manchester
Teaching
Former lecturer in digital marketing, Manchester Metropolitan University
Industry roles
Search awards judge and conference programming committee member
Known for
Explaining BERT, information retrieval, and RAG to marketers
Dawn Anderson is the founder of Bertey, a Manchester SEO consultancy, and a search marketer best known for translating information retrieval research into plain advice for working SEOs. She has taught digital marketing at Manchester Metropolitan University and spent years on the judging and programming side of the industry, at search awards and on conference committees.

The Explainer of BERT

When Google brought BERT into search in 2019, many marketers needed someone to explain what a language model for search really did. She wrote A deep dive into BERT: How BERT launched a rocket into natural language understanding, and spoke on Google BERT at SMX West, walking marketers through natural language understanding, context, and why it mattered for queries.

Her consultancy takes its name from the same model. That background, years of reading retrieval and ranking papers and putting them into language marketers can use, is the lens she now brings to AI search. Our GEO and AEO industry statistics chart how fast that topic grew.

You Can't Generate What You Can't Retrieve

In January 2026 she published a talk with a title that works as a thesis: You Can't Generate What You Can't Retrieve. Large language models, she argued, depend on retrieval systems to produce grounded answers, and Retrieval-Augmented Generation pipelines reintroduce classic search problems under a new name.

Document structure and chunking, relevance, freshness, and evaluation are, in her view, the same decisions SEOs have always influenced. Now they directly affect what AI systems seem to know. Her slides close on a short list: be explicit for clarity, stay fresh, and design for retrieval.

The Retrieval Bottleneck

At the SEO and GEO Summit in March 2026 she went further in The Retrieval Bottleneck, subtitled Why Your AI Strategy is Still a Search Strategy. The industry, she said, had spent a year obsessing over model parameters and prompt engineering, while the biggest point of failure in generative AI remained information retrieval.

When AI answers a question wrongly, she argued, it is rarely a hallucination problem. It is usually a structure, freshness, or relevance problem. Her new definition of visibility fits on one slide: “Being retrievable, understandable, and executable by machines.” Our AI crawler blocking statistics show one way sites still make themselves impossible to retrieve.

From Ranking First to Being Retrievable

Her mandate for SEOs is a shift in goal, from ranking first to what she calls contextual retrievability. Document hierarchy and metadata, she says, now serve as the main map AI models use to find and lift content.

It is the same message she has given since BERT, aimed at a new audience: understand how the machine reads before you try to influence it. For marketers buying GEO services, our guide to GEO agencies is a useful place to test which firms talk about retrieval instead of tricks.

Links

← All people

Sources