The AI Persuasiveness Analyst skill for Claude

You recognise the problem the skill solves by the question you ask yourself after hitting send: why is nobody replying? The text was neat, the spelling was correct, the offer was good. But neat is not persuasive. Possible gaps to inspect alongside language, offer and context include: there is no proof in it that others already trust you, there is no reason to reply now rather than next month, and the core message is buried in paragraph three. Those are not style errors, they are missing principles, and you can only find them by looking systematically.

That systematic approach is the core of the file. It runs through Cialdini's seven principles, the six pillars of Made to Stick and Kahneman's layers of thinking, and ticks off per principle whether it is demonstrably in the text. Demonstrably means here, literally: with a direct quote from your text as proof, or a short explanation of why it is missing. By its own description it works for landing pages, cold emails, sales emails and nurture flows, adverts on Meta, Google and LinkedIn, pitches and one pagers, whitepaper introductions, social posts with a commercial goal, and quotes.

It belongs in the skill library that we make freely available from the AI and automation service, with no account and no sales email afterwards. For marketers, copywriters and business owners who do not want to guess why a text does or does not work.

There is proof, but of the wrong kind. The email leans entirely on authority: ten years of experience, ISO certified. That is one of the seven principles, and it is the principle that is about you. Social proof, the principle that is about people like the reader, is completely missing. Relevant evidence about similar customers may help, but the text alone does not establish how this audience will respond.

Nothing sticks. Held up against the six pillars of Made to Stick, the email falls flat: the core message only appears in paragraph three, the opening is predictable, the promise of more results is not concrete, and there is no emotion and no story in it. The Heath brothers keep asking the same question: does this stick or not. For this email, missing features are editorial observations; recall has not been measured.

There is no reason to reply now. The Kahneman layer looks at what the text does to fast, intuitive thinking. Here: not much. No anchor value that colours the rest of the numbers, no loss framing. The email proposes a growth benefit but no explicit loss frame. Prospect Theory concerns choices under risk; it does not show that loss-framed copy universally converts better. A relevant, truthful benefit can be appropriate without manufactured urgency.

The revised example does not invent competitors or growth percentages. It asks a reader-centred question using only the supplied text. The full download includes three rewrites, with any unverified certification detail marked as an unfinished field rather than publication-ready proof.

After that comes the process, seven steps in fixed order. Read the text in full. Classify the goal: inform, sell, recruit or persuade a decision. Determine the audience and the use case if they are supplied, and ask if they are not. Run through the three frameworks in order: Cialdini, Heath, Kahneman. Apply the stated 15-criterion rubric, show ratings and weights, and normalise to 100. Write the rewrites with the weakest part first. And close with a final conclusion in two sentences.

Observed features require verbatim quotations; absent elements require a scope-based explanation, not an invented quote. Do not fabricate customer relationships, results, certifications or urgency in a rewrite. The seven analysis rules also limit style advice, refer back to the sources, require a proposed rewrite, use British English unless the source text is Dutch, and spell AI in capitals.

The file closes with a worked example of a full analysis, the report you saw shortened in FIG.01, and a source list of four titles: Cialdini, the Heath brothers, Kahneman, and the original Prospect Theory paper by Tversky and Kahneman. If you want to learn to build an instruction file like this yourself, the full method is in writing SKILL.md NL.

The influence comes from Robert Cialdini. The expanded 2021 edition of Influence is used, with seven principles: reciprocity, consistency, social proof, liking, authority, scarcity and unity. Each principle identifies a possible influence to examine, and the check is simple: is that route demonstrably in your text or not. Note what the skill does not do: it does not teach you to apply the principles in a conversation or campaign, it tests whether they are in a text.

The stickiness comes from Chip and Dan Heath. Made to Stick from 2007 answers a different question than Cialdini: not why people say yes, but why some messages stick and others evaporate. The six pillars, from Simple to Stories, form the second block of every report.

The thinking comes from Daniel Kahneman. Thinking, Fast and Slow from 2011 describes the two layers in which people think: the fast, intuitive System 1 and the slow, rational System 2. Deliberate reading can require effort, but effort is not universally harmful to conversion and the text does not reveal a reader’s actual cognitive process. That is why it checks for the shortcuts that do speak to System 1: anchoring, framing, availability and familiar words.

The loss framing comes from Tversky and Kahneman together. Prospect Theory from 1979 models gains and losses relative to a reference point in decisions under risk. It does not validate a universal advertising effect. Inspect truthful, relevant loss framing without treating its absence as an automatic defect or inventing urgency.

If you want to work on the structure of the argument after the analysis, you can continue with two neighbours from the library: the Pyramid Principle Writer skill for the order of the argument, and the SCQA Communication Coach skill for the tension arc of your opening. This analyst tells you what is missing; those two help you fill the gap.

After that, using it is simple: paste the text and ask why it does or does not work. Add the goal and the audience if you know them, because the weighting of the score depends on it; if they are missing, the skill asks for them. If you are new to files like this, the knowledge base has the wider explanation of working with AI.

There are also limits. You will not get style advice here: spelling, rhythm and tone are editing work, and that is not what you come to a psychological analyst for. It does not predict conversion: the score summarises explicit editorial evidence ratings and relevance weights, not how many people will click tomorrow, because that also depends on your offer, your price and your audience. And it does not turn a bad offer into a good one. If nobody wants your service, that is not down to a lack of loss framing.

The most honest limit sits in how you use the outcome. What gets exposed here are persuasion levers, and you can pull those the wrong way too: inventing scarcity that is not there, claiming proof you do not have. Do not do that, and the skill will not help you with it: it points out where a principle is missing, but the real proof has to come from your own records. The texts before and after in the chain have their own skills: the SPIN Sales Conversation NL for the conversation where you gather the arguments, and the Negotiator for the conversation that follows your persuasive proposal.

A more persuasive text raises the odds that a reader acts, but there have to be readers in the first place. If too little traffic reaches your pages, the real work ahead of this analysis is: being found by the people you are looking for. The free SEO scan shows in a few seconds where your site stands. In the guide writing models compared you can see how this model relates to the others. And if you want to talk through what else AI can do for your business, from single skills to full automation, we simply do that in a conversation.