Below is a compact, practical guide + ready-to-use prompt templates you can plug into ChatGPT (or other LLM) to generate product pages, category pages and CMS pages for PrestaShop. The templates are designed to produce SEO-ready text, short & long descriptions, meta tags, image alt text, attributes & specs, and optional CSV/JSON output for import. Key principles (what makes a “perfect” prompt) - Give the model a clear role and goal (e.g., “You are an expert e‑commerce copywriter for PrestaShop”). - Provide structured input data (fields) and tell the model exactly which outputs you want. - Specify tone, audience, length, and SEO focus (primary keywords). - Ask for multiple formats/variants (short, long, bullets, meta, social copy). - Request machine‑readable output (JSON or CSV) if you plan to import to PrestaShop. - Include constraints (character limits for meta title/desc, no HTML or only safe HTML tags, language). - Ask for editorial safeguards: no trademark misuse, factual checks, or “do not invent specs” unless allowed. - Iterate: ask for 3 variations, then refine. Use delimiters around input data. General system + user prompt pattern - System: “You are an expert e‑commerce copywriter and SEO specialist who writes product descriptions, category descriptions and CMS page copy optimized for PrestaShop imports.” - User: Provide product/category/CMS data (structured) and a clear instruction block that asks for outputs and formats. Templates 1) Product prompt (single product) System (optional): You are an expert e‑commerce copywriter and SEO specialist for PrestaShop. Produce copy that converts, follows brand voice, and is ready for import. User: Use this exact structure. Output JSON only (no extra commentary). Observing these constraints: - Meta title: 55–60 characters max. - Meta description: 120–155 characters. - Short description: plain text, 1–3 sentences, 150–300 chars. - Full description: HTML allowed (

,

,
• , ) or plain text if requested. - Provide 3 headline/title variants, 3 short-description variants, and 3 full-description variants. - Provide image alt texts (one per image) and 5 product tags. - Provide product features as bullets and as a JSON list of {name, value}. - Provide 5 SEO keywords / search phrases (priority order). - Do not invent technical specs unless I marked them as “confirmed”. Input data (replace placeholders): { "product_name": "{{product_name}}", "brand": "{{brand}}", "category": "{{category_name}}", "condition": "{{new|used}}", "short_spec": "{{one-line spec}}", "features": ["feature1", "feature2", "..."], "materials": "{{materials}}", "colors": ["color1","color2"], "sizes": ["S","M","L"] or combinations, "target_audience": "{{e.g., men 25-45, outdoors}}", "primary_keyword": "{{primary keyword}}", "secondary_keywords": ["kw2","kw3"], "images": ["img1.jpg","img2.jpg"], "brand_voice": "{{friendly|luxury|technical|playful}}", "language": "{{en|fr|de}}", "price": "{{price}}", "warranty": "{{warranty info or blank}}", "confirmed_specs": {optional JSON of specs you want the model to treat as factual} } Required outputs (JSON keys). Example: { "name_variants": ["...","...","..."], "meta_title": "...", "meta_description": "...", "friendly_url": "...", "short_descriptions": ["...", "...", "..."], "full_descriptions_html": ["

  ...

  ", "

  ...

  ", "

  ...

  "], "features_bullets": ["...","..."], "specs": [{"name":"Weight","value":"500 g"}, ...], "image_alts": ["alt for img1","alt for img2"], "seo_keywords": ["...","...","...","...","..."], "tags": ["tag1","tag2","tag3","tag4","tag5"], "suggested_category": "...", "cta": "Buy now • Free shipping • etc." } Example (concrete filled prompt snippet) Input data: { "product_name": "AeroTrail Waterproof Hiking Jacket", "brand": "PeakPath", "category": "Men > Outdoor Clothing > Jackets", "condition": "new", "short_spec": "3-layer waterproof, breathable membrane; 10,000 mm WP; taped seams", "features": ["waterproof 10,000 mm", "breathable membrane", "adjustable hood", "zip hand pockets"], "materials": "Nylon + PU membrane", "colors": ["navy","green"], "sizes": ["S","M","L","XL"], "target_audience": "men 18-45 who hike and camp", "primary_keyword": "waterproof hiking jacket", "secondary_keywords": ["men's hiking jacket","breathable outdoor jacket"], "images": ["aerotrail-front.jpg","aerotrail-back.jpg"], "brand_voice": "practical friendly", "language": "en", "price": "129.99", "warranty": "2 years manufacturer warranty" } (Then the model returns the JSON described above.) Notes/variations: - If you prefer CSV, ask for a single-line CSV output with columns: name,meta_title,meta_desc,short_desc,full_desc,friendly_url,tags,image_alts,sku,price,category. - For combinations (variants like size/color) ask the model to output combination SKUs and attribute values as an array. 2) Category prompt (category description + SEO) System: same role. User: Input: { "category_name": "{{category_name}}", "parent_category": "{{parent}}", "audience": "{{audience}}", "primary_keyword": "{{primary_keyword}}", "secondary_keywords": ["..."], "product_examples": ["product1","product2"] optional, "brand_voice": "{{voice}}", "language": "{{lang}}", "page_length": "short|medium|long" (short=80-150, medium=160-300, long=300-700 words) } Required outputs: - meta_title (50–60 chars) - meta_description (140–155 chars) - friendly_url (slug) - 2-3 headline variants - category_intro (one paragraph) - full_category_description (SEO optimized to length requested; include primary keyword 2–4 times, secondary 1–2 times; include 2 internal link suggestions) - recommended subcategories or filters (bulleted) - suggested banners/campaign copy (short) 3) CMS page prompt (e.g., About us, Shipping policy, FAQ) System: same. User: Input: { "page_type": "About us | Shipping | Returns | FAQ | Size guide | Landing campaign", "company_name": "Brand", "tone": "friendly|professional|luxury", "required_sections": ["Our story","Sustainability","Team","Contact"], "language": "en", "target_length": "short|medium|long", "legal_requirements": "refund policy details to include", "primary_keyword": "about brand name" } Required outputs: - page_title - meta_title & meta_description - full page HTML (using safe tags:

  ,

  ,

  ,

  ,
  • , ) - 3 suggested FAQs (question + short answer) - internal links to suggest (e.g., contact, shipping, category pages) - CTA (email signup, contact form link, etc.) Examples of short prompts you can use directly - Product (short): “You are an e-commerce copywriter. Rewrite the following product description for PrestaShop product import. Output: meta_title (<=60 chars), meta_description (<=155 chars), short_desc (150–220 chars), full_desc HTML. Product: [paste fields]. Tone: modern and energetic. Primary keyword: [kw]. Return JSON only.” - Category: “Write an SEO-optimized category page for [Category name] aimed at [audience]. Include meta title, meta description, friendly URL, 250–350 word description with primary keyword [kw] used 3 times. Output plain text.” How to handle multilingual content - Provide the language and ask for translations. Prefer using the same prompt but add "Return translations for: en, fr, de" and request separate JSON objects for each language. - Ask for localized idioms and avoid literal machine translations of brand terms. Formatting for PrestaShop import - If you import via CSV, set the CSV header order to pre-stated PrestaShop fields. Ask the model: “Return a single CSV line with these columns: name,meta_title,meta_description,description_short,description,link_rewrite,tags,image_1_alt,...” - If you use an API or module that accepts JSON, map to that module’s schema and ask for that exact JSON. Advanced tips and guardrails - Character limits: always state exact character limits for meta fields. PrestaShop typically shows meta titles ~55–60 chars and meta descriptions ~120–160, but that’s flexible — set exact numbers you want. - Ask for “no brand claims” or “do not claim medical benefits” if sensitive. - For technical products, request a separate “technical spec table” and ensure the model only uses confirmed specs. - Provide competitors or example product pages to match tone and format. - Use “temperature” lower (0.0–0.4) for consistent, factual copy; higher for creative variations. - Request multiple variations (A/B test) and short social/ad snippets (30–90 chars). Example outputs to request for each product (useful checklist) - 3 product title variants (50–70 chars) - short description (150–220 chars) - long HTML description (200–800 words or chars as needed) - meta title, meta description, friendly URL - image alt text - 5 tags and 5 SEO keywords - 3 short ad headlines for social channels - 5 FAQ lines relevant to the product - product attributes/specs and combination SKUs Quick ready-to-copy product prompt (one-liner) You are an expert e-commerce copywriter and SEO specialist. Given this JSON: {product_name,brand,category,short_spec,features,materials,colors,sizes,primary_keyword,secondary_keywords,images,brand_voice,language,price,warranty}, output JSON only with keys: name_variants[3], meta_title(<=60), meta_description(<=155), friendly_url, short_descriptions[3], full_descriptions_html[3], features_bullets, specs(list), image_alts(list), seo_keywords(5), tags(5), cta. Do not invent specs not in input. Final note Start with a small batch (10–20 products) to refine the prompt and ensure outputs match your PrestaShop fields and s. Then scale and automate (via API) once you’ve validated the structure and copy. If you want, paste one product, one category, or one CMS page data now and I’ll generate an exact prompt + sample output ready for PrestaShop import.