Where your customers’ own words are kept and how to collect them without changing a word. The lesson explains buyer intent, the question a buyer is trying to answer when they arrive. It then covers the six places customer words are kept: reviews, support emails and chats, site search terms, return reasons, pre-sale questions and competitors’ reviews, with where to export each one. You learn the five collection steps, including removing personal data before anything goes into an AI tool, and why each question is tagged rather than each customer. Every quote is saved word for word as one row in customer-voice.md, with a ref, its source and the product it is about. Terms explained include buyer intent, verbatim, pain point, objection and persona.
How to tag every quote in your customer voice file and turn the quotes into a ranked list of what your buyers struggle with. Each quote gets five tags. Its kind is pain, desire, objection, trigger or praise, and its theme is what it is about. Its emotion is frustration, confusion, fear, hope or delight. Its intensity is mild, strong or severe, and its buying stage is awareness, commercial or transactional. You learn how Claude adds these tags without touching the customer’s words and why a tag must point to a word in the quote. You then group the rows by theme, count and rank them, and keep the leftovers in a counted "other" group. The result is pain-dictionary.md, with each theme’s count, emotion, stage and the clearest quote word for word.
How to group your buyers into three predictive personas and lay every theme out by buying stage and emotion. The lesson compares demographic personas with predictive ones built from what buyers say. It walks through the ten persona fields, from pain, trigger and risk filter to coverage and evidence. You learn how to group quotes by shared trigger and desire, count how much of the file each persona explains, and keep a named residual group. If that group is over 20%, a fourth persona is missing. The intent matrix places every quote in one of fifteen cells, three buying stages against five emotions, and each full cell names the page, ad or email it needs. You finish with personas.md and intent-matrix.md, and see how every paid e-commerce course reads them.
The six prompts from the SOP & AI Prompts pack, run in order in one Claude conversation on real reviews, support questions and site searches. They build the customer voice file, check every quote word for word against the exports, add the five tags, and produce the pain dictionary, three personas and the intent matrix. You leave with all four files saved and ready to load into a paid course.