Definition
What is the Botworthy Score?
The Botworthy Score is a 0 to 100 score for how likely AI models are to quote a page, across 10 criteria weighted by published research into what AI models actually cite.
The Botworthy Score was created by Jessica Redman, who coined the name in 2026. It began as a rubric built to score a major UK insurer’s home insurance knowledge centre, where the average article scored 45.6 out of 100 and 43% of sections failed to answer their heading in the first sentence. It is published by Botworthy, a service of Work That Limited.
How the Botworthy Score is calculated
Every page is scored out of 100 across four areas. Each area asks one question of the page and holds several criteria. Each criterion has a floor, where it scores zero, and a target, where it scores full marks, with a straight line in between.
| Area | Points | What it asks | Example criteria |
|---|---|---|---|
| Answerability | 33 | Does each section answer its heading straight away? | Answer-first sections, question-led headings, a summary at the top |
| Extractability | 33 | Can a passage be lifted out and still make sense? | Self-contained passages, short sentences, lists and tables |
| Evidence | 26 | Is there something worth citing? | Statistics and specifics, named sources, named expertise |
| Freshness | 8 | Has it been updated recently? | Date of the last substantive update |
The weights are set by published research into what AI models cite: the Princeton GEO paper (Aggarwal et al., KDD 2024), Cyrus Shepard’s Zyppy meta-analysis of 54 studies, and Ahrefs’ analysis of almost 17 million AI citations. Where the evidence for a criterion is strong it carries more weight; where it is thin it carries less. Answer-first writing is the single biggest factor.
Free and live versions
The free version is a Claude skill on GitHub, built from a fixed snapshot of three studies. Its weights do not change. The live version is re-weighted every month as new citation research is published, and is part of Botworthy membership at £20 a month plus VAT, with a research brief, a scoreboard showing the evidence behind every weight, and the current skill.
How to cite the Botworthy Score
Redman, J. (2026). The Botworthy Score: a research-weighted score for how likely AI models are to quote a page. Botworthy. https://botworthyindex.com/botworthy-score
The method behind the first study, including the statistical checks, is written up in The scientific scoring rubric I use to analyse whether content is GEO optimised.
Questions
What is the Botworthy Score?
The Botworthy Score is a 0 to 100 score for how likely AI models are to quote a page, across 10 criteria weighted by published research into what AI models actually cite. A high score means a page is written so that a model can lift a passage out and quote it; a low score means the answer is buried, vague or unsupported.
Who created the Botworthy Score?
Jessica Redman, a UK GEO, SEO and AI enablement consultant, created the Botworthy Score and coined the name in 2026. It began as a rubric she built to score a major UK insurer's knowledge centre, and is published by Botworthy, a service of Work That Limited.
How is the Botworthy Score calculated?
Each page is scored out of 100 across four areas: answerability (33 points), extractability (33), evidence (26) and freshness (8). Each area holds criteria with a floor and a target, scored on a straight line between them. The weights are set by published citation research, including the Princeton GEO paper (KDD 2024), Zyppy's meta-analysis of AI citation factors and Ahrefs' analysis of almost 17 million AI citations.
Why is it called Botworthy?
Because the reader of a page is now, very often, a bot. The score asks whether a page is worth a bot's quote, not only whether it is worth a person's time.
Is the Botworthy Score free?
The free version is a Claude skill on GitHub, built from a fixed snapshot of three studies. The live version, which is re-weighted as new research is published, is part of Botworthy membership at £20 a month plus VAT.
Does a high score guarantee AI citations?
No. The score measures what is on the page. Whether a page can be crawled, how it ranks in search and how often the brand is mentioned elsewhere matter too. In the first study, the highest-scoring articles were also the ones AI models cited most.