Botworthy™

Is your content Botworthy?

AI models answer your customers’ questions by quoting someone’s content. The Botworthy Score measures how likely they are to quote yours, across 10 criteria weighted by published citation research.

Get the free scoring skillSee the evidence
Illustrative example, showing some of the criteria we score against. Each row is an article; darker means more of it met. A column that stays pale all the way down is a habit across the whole library.

AI models quote your first sentence. Make it the answer.

When a model builds an answer, it lifts passages that make sense on their own. If your answer sits in sentence three, behind a warm-up line or an ‘it depends’, it rarely gets lifted.

Answer-first writing is the single biggest factor in the Botworthy Score.

How often should a boiler be serviced?

Your boiler works hard all year, especially through the colder months. It depends on the type of boiler you have and what your warranty says, but most engineers recommend an annual service.

What an AI model lifts

“Your boiler works hard all year, especially through the colder months.”

Doesn't answer the question

The answer is buried behind a warm-up line and an 'it depends'.

Illustrative example.

The Botworthy Score

Every article is scored out of 100 across four areas: answerability, extractability, evidence and freshness. Each area holds several criteria, and each criterion carries a weight set by published research into what AI models actually cite. When the research changes, the weights change with it.

Current version

v3.0.4, October 2026

Every weight is anchored to published research into what AI models cite, about ninety studies at the last count. When a new study changes the picture, the weight moves, the change is written up, and the skill is sent out again. The latest change: v3.0.4: two watch metrics, no weight change.

See every weight and the research behind it

Free

Run it on your own content

The score runs as a skill in Claude. Leave your email and the current version arrives within a minute, with the install steps. You hear from us when the score changes, and not otherwise.

Get the skill by email

The writing habits that cost citations

Scoring every section opening shows the same few habits again and again. Writers pick them up from years of writing for the web, which is why nobody spots them in their own work. Score a whole library and the habits show up as patterns across your content team, not one-off mistakes. From there you can build the training and templates that re-teach the team, and rescore to check it stuck.

The unanswered question

A heading written as a question, followed by a paragraph that never answers it.

Write instead: Answer it in the first sentence, then explain.

The 'it depends' opener

A section that opens with 'it depends' or 'that's up to you' and leaves the condition for later.

Write instead: Give the usual answer and the condition together.

The announced list

“Here are a few things to consider:”, which tells the reader a list is coming without saying what's in it.

Write instead: Summarise the list in one sentence, then give it.

The warm-up line

A general or emotional sentence before the point, such as “We all know how stressful moving can be.”

Write instead: Delete it. Your second sentence is usually the right first one.

Why before what

An instruction heading that opens with why it matters instead of what to do.

Write instead: Instruction first, reason second.

Get the Botworthy skill

The score is free. It is built on published research, every weight is public, and every change is explained. In return we ask for your email, so the new version reaches you when the evidence moves.

What you get

  • The scoring skill, current versionThe Botworthy Score runs as a skill in Claude. It scores 25 to 40 articles at a time, codes every section opening twice, finds the writing habits behind the weak ones and writes the report. You get the version that matches the evidence today.
  • The scoreboardEvery criterion, the weight it carries, how it is measured and the studies behind it, with a changelog whenever a weight moves. It is public, so you can check the reasoning before you trust the score.
  • A post every time the score changesNew research on AI citations comes out every month. When it changes a weight or a rule, we publish what moved, by how much and which study moved it, and email you the new skill.
  • Patterns across your team, not just scoresThe skill groups weak sections by the writing habit behind them, so you can see which habits cost citations across your whole content team, build the training to fix them, and rescore to prove it worked.
  • Index data as it landsSector scores from each Botworthy Index study, so you can benchmark your content against the leading brands in your market.

The method was first run on a FTSE 100 company’s content library, where the highest-scoring articles were also the most cited by AI models.

The Botworthy Index

We’re scoring the content of the leading brands in one sector at a time and publishing the rankings. Get the first Index in your inbox when it’s out.

Who’s behind it

Botworthy was created by Jess Redman, who has spent eight years in SEO and digital marketing, in-house and for clients in insurance, pharma, wellness and IoT. She’s spoken at BrightonSEO and DMWF, and in Lisbon on how to get content cited by AI models.

She built the Botworthy method to show content teams exactly why AI models pass over their pages, and what to change. Follow Jess on LinkedIn.

Questions

Is the Botworthy Score free?

Yes. The current version of the scoring skill is free in exchange for your email address, which is how you hear when the score changes. There is no card and no account.

How often does the score change?

Whenever the evidence does. Each change is a numbered release with a post explaining what moved and why, and the skill is sent out again. Nothing moves on the strength of one study or one website.

What do I need to run it?

Claude, and Python 3 with numpy and scipy if you run it in Claude Code. Save the skill file to your commands folder and run it against a section of your site. The install steps come with the email.

Does a high score guarantee AI citations?

No. The score measures how quotable a page is once an AI model has found it. Retrieval rank, crawlability and mentions on other sites matter more, and the scoreboard says so. In the first study, the highest-scoring articles were also the ones AI models cited most.