, ) - tone: {TONE} (e.g., friendly professional) Return: - JSON object with keys: name, sku, brand, category, price, short_description, description, meta_title, meta_description, friendly_url, tags (array), images (array of {filename, alt_text}), features (array of {name, value}), attributes (array of {group, value, impact_price, quantity}), weight, dimensions, ean13, upc, quantity, available_for_order, condition. Do NOT include explanations or markdown — only the JSON. B) Bulk CSV generator for many products Prompt: You are generating a CSV for PrestaShop import. Create rows for these products. Use header row exactly: "name","sku","category","price","short_description","description","meta_title","meta_description","friendly_url","tags","images","features","attributes","quantity","weight","ean13" For each product, fill fields. Field rules: - description: plain HTML allowed; short_description should be plain text and ≤160 chars. - images: semicolon-separated filenames (no URLs). - features: semi-colon separated feature:value pairs. - attributes: semicolon-separated groups in this format Group=Value|impact_price|quantity (e.g., Color=Red|0.00|10;Size=M|1.00|5) Return the full CSV content only, no commentary. C) Category description prompt Prompt: Write a category page for PrestaShop: - category_name: {CATEGORY_NAME} - parent: {PARENT_CATEGORY} - target_audience: {AUDIENCE} - SEO primary keyword: {KEYWORD} Constraints: - meta_title ≤ 70 characters, meta_description ≤ 160 characters - category_description: 2 short paragraphs (max 400 words), include the primary keyword once in the first paragraph and once in the meta_description. - Provide three recommended subcategory names. Return JSON with keys: name, parent, meta_title, meta_description, friendly_url, description, recommended_subcategories (array). D) CMS page (About / FAQ / Policy) Prompt: Write a PrestaShop CMS page: {PAGE_TYPE} (About / Shipping policy / Returns / FAQ) Inputs: - company_name: {COMPANY_NAME} - tone: {TONE} - target_customers: {AUDIENCE} Constraints: - meta_title ≤ 70, meta_description ≤ 160 - Include headings (H2) for sections and a short intro paragraph. - For FAQ: produce 7 Q&A pairs. Return: JSON keys title, meta_title, meta_description, friendly_url, content (HTML with H2/H3 and
), faqs (if any). E) Generate image alt text and SEO tags for product images Prompt: Given product: {PRODUCT_NAME} + {SHORT_DESC}. For images: {image1.jpg,image2.jpg,...} Provide alt text for each image (max 100 chars), and a list of 10 SEO tags/keywords prioritized for search relevance. Return JSON: {images: [{filename, alt_text}], seo_tags: []}. 4) Example prompt + example output (short) Prompt (single product): You are an e-commerce writer for “GreenHome” (eco-lifes brand). Create a PrestaShop-ready JSON product record only. Inputs: - product_name: Bamboo Cutting Board – Large - brand: GreenHome - category: Kitchen > Cutting Boards - SKU: GH-CB-L-01 - price: 29.99 USD - short_specs: bamboo, non-slip feet, juice groove, 45×30×2 cm, 1.2 kg - keywords: bamboo cutting board, eco cutting board, sustainable kitchen - tone: friendly professional Follow the constraints from template A. Example output (trimmed): { "name":"Bamboo Cutting Board – Large", "sku":"GH-CB-L-01", "brand":"GreenHome", "category":"Kitchen > Cutting Boards", "price":"29.99 USD", "short_description":"Large sustainable bamboo cutting board with non‑slip feet and juice groove.", "description":"
Our GreenHome Bamboo Cutting Board is crafted from sustainably sourced bamboo for a durable, knife-friendly surface...
- Size: 45×30×2 cm
- Weight: 1.2 kg
- Non-slip rubber feet
- Juice groove
", "meta_title":"Bamboo Cutting Board – Large | GreenHome", "meta_description":"Large eco-friendly bamboo cutting board with juice groove and non-slip feet. Durable, knife-friendly. Free shipping over $50.", "friendly_url":"bamboo-cutting-board-large", "tags":["bamboo cutting board","sustainable kitchen","eco cutting board"], "images":[{"filename":"gh-cb-l-01-1.jpg","alt_text":"Large bamboo cutting board with juice groove"},{"filename":"gh-cb-l-01-2.jpg","alt_text":"GreenHome bamboo board non-slip feet"}], "features":[{"name":"Material","value":"Bamboo"},{"name":"Size","value":"45×30×2 cm"},{"name":"Weight","value":"1.2 kg"}], "attributes":[{"group":"Side","value":"Reversible","impact_price":"0.00","quantity":"100"}], "quantity":150, "weight":"1.2", "ean13":"1234567890123" } 5) Tips for accuracy, SEO and quality - Supply as much product data as possible (materials, sizes, specs, compatibility). The model will hallucinate if data is missing. - Provide primary + 2 secondary keywords and specify where to place them. - Limit characters for meta fields; request a word/char count. - Ask for HTML-safe content if you will paste into WYSIWYG. - Request multiple meta variants or short/long descriptions for different channels. - For large catalogs, produce CSV/JSON chunks (e.g., 50–200 rows) to avoid token limits. - Validate EAN/UPC and legal claims separately; don’t rely on AI for compliance-critical facts. 6) Prompts for combinations and attributes - Provide a clear mapping: attribute group (Color) -> values (Red, Blue) -> SKU suffix -> stock -> price impact. Example short instruction: Create combinations for Color:Red|Blue and Size:S|M|L. Base SKU GH-TSH; SKU format GH-TSH-{ColorCode}-{Size}. Price impact: Size L +2.00, Colors no impact. Output JSON array of combinations with sku, attributes, price, quantity. 7) Iteration and quality checks - First run: ask for 3 variants (conservative, promotional, SEO-focused). Pick one and ask the model to rewrite to match chosen variant exactly. - Use explicit negative instructions: “Do not use phrases like ‘best on the market’ or ‘number one’.” - Compare generated friendly_url, meta_title, meta_description for duplicates across catalog. - Spot-check for hallucinations: ask the model to list which fields were inferred vs provided. 8) Quick checklist to include in every prompt - Brand name, tone, audience - Product name and SKU - Category path - Price + currency - All technical specs (size, weight, materials, compatibility, EAN/UPC) - Primary SEO keyword + 2 secondaries - Required output format (JSON/CSV) and exact field names - Limits for meta fields and word counts - Request “only the data” (no commentary) if you need import-ready output 9) Example full prompt you can paste into ChatGPT You are an e-commerce content engine for our PrestaShop store. Produce a JSON array of 10 product objects ready for import. Each object must include these keys: name, sku, category, price, short_description, description (HTML), meta_title, meta_description, friendly_url, tags (array), images (array of {filename,alt_text}), features (array), attributes (array), quantity, weight, ean13. Use the following raw product input lines (give me the products below): [paste CSV-s list of product inputs]. Constraints: meta_title ≤70 chars; meta_description ≤160; friendly_url lowercase with hyphens; descriptions must not mention stock levels or delivery times. Return ONLY the JSON array. 10) Final notes - Match the PrestaShop field names your import tool expects (some use different headers). If you use PrestaShop Manager, adapt header names accordingly. - For translation/multi-language stores, ask the model to output language keys (en, fr, es) or produce separate files per language. - Automate by building a prompt template in your tool and replacing placeholders with product fields from your database. If you want, paste a real product (or a small CSV of several products) and I’ll generate a ready-to-import JSON or CSV file based on the templates above.
