Below is a compact guide plus ready-to-use prompt templates you can copy/paste into ChatGPT (or another LLM) to reliably create product pages, category pages and CMS pages for PrestaShop. Each template tells the model its role, the inputs, the output format (JSON/CSV/HTML), constraints (length, SEO limits), and failure-handling rules so you get import-ready content.
General best practices before prompts
- Always include concrete product data (brand, SKU, specs, materials, colors, dimensions, images, price, availability). If the model must guess, tell it to ask instead or mark fields "MISSING".
- Use a clear output format (JSON or CSV) so you can import or parse automatically.
- For consistent tone and terminology, set role + tone (e.g., “You are a senior e‑commerce copywriter for premium home goods”).
- Enforce SEO limits: meta title ≤ 60 chars, meta description ≤ 155–160 chars, slug friendly (lowercase, hyphens).
- Keep temperature low (0.0–0.4) for deterministic outputs for imports; use higher when brainstorming.
- Ask for multiple variants (2–3 versions) if you want A/B testing.
- Validate outputs: check product facts (sizes, ingredients) — LLMs can hallucinate specs.
Templates
1) Single-product, full package (short + long + SEO + alt tags + JSON-LD)
Prompt:
You are a senior e‑commerce copywriter specialized in PrestaShop product content. Create import-ready output in JSON with these keys: name, brand, reference, sku, category, price, short_description_html, description_html, meta_title, meta_description, link_rewrite (slug), tags (comma-separated), image_alts (list), attributes (JSON array of {name,value}), technical_specifications (HTML table), jsonld (Product schema JSON‑LD). Follow these rules:
- short_description_html: max 200 characters (HTML allowed), one short marketing sentence.
- description_html: between 200 and 800 words, use H2/H3 sections, bullet lists for features, benefits, how-to-use, materials, care.
- meta_title ≤ 60 chars; meta_description ≤ 155 chars.
- link_rewrite: lowercase, hyphens, no stopwords, max 80 chars.
- image_alts: generate one alt text per image filename provided.
- If any required field is missing, set it to "MISSING - ask for info".
- Do not invent technical specs; if missing, put "MISSING".
Input data:
Name: [Product name]
Brand: [Brand]
Reference/SKU: [ref/sku]
Category: [Category path]
Price: [price currency]
Primary features (comma list): [features]
Technical specifications: [list or “none”]
Images: [image1.jpg, image2.jpg,...]
Target audience: [audience]
Primary keywords: [kw1,kw2]
Tone: [friendly/professional/luxury]
Output only valid JSON.
Example usage (fill placeholders) — model should produce a JSON object ready to map into PrestaShop fields and a jsonld string you can paste into product template.
2) Bulk CSV generator for multiple products (PrestaShop import-ready)
Prompt:
You are an automated PrestaShop CSV generator. Produce a CSV with the following headers (in this exact order): id, active, name, category, price, wholesale_price, tax_rules_group_id, quantity, minimal_quantity, description_short, description, meta_title, meta_description, link_rewrite, reference, supplier_reference, sku, ean13, upc, weight, width, height, depth, image_urls, tags, visibility, condition. Use the input rows below and fill missing fields with "MISSING". Ensure description_short and description are HTML-friendly (use
/- for bullet points). For meta_title and meta_description enforce ≤ 60 and ≤ 155 characters respectively. Output only CSV content, no extra commentary.
Input rows:
1) Name: [..], Brand: [..], Category: [..], Price: [..], Quantity: [..], Features: [..], Specs: [..], Images: [url1;url2], Keywords: [..]
2) ...
(Include all rows you want processed)
3) Category page content and SEO
Prompt:
You are an SEO copywriter creating a PrestaShop category page. Output JSON with: category_name, parent_category, description_html, meta_title, meta_description, link_rewrite, image_alt, suggested_subcategories (array of up to 6). Rules:
- description_html: 150–400 words, include 2–3 keyword variations naturally, one H2, an intro paragraph, a product-filtering tips section, and a short CTA line.
- meta_title ≤ 60 chars; meta_description ≤ 155 chars.
- link_rewrite: friendly slug.
Input:
Category name: [name]
Parent category: [parent]
Primary keywords: [kw1,kw2]
Tone: [tone]
Top products included (optional): [product list]
4) CMS page (About Us, Shipping, Returns) with sections + FAQ + schema
Prompt:
You are a content specialist writing PrestaShop CMS pages. For each page requested, produce HTML content with H1, H2, paragraphs, bullet lists, FAQ section (5 Q&A), and a JSON‑LD FAQ schema. Also produce meta_title (≤60) and meta_description (≤155). Follow the brand voice: [voice brief]. For policy-type pages (shipping/returns), include timeframe, costs, contact info placeholder, steps, and a clear CTA for buyers. Output JSON with keys: slug, title, html_content, meta_title, meta_description, jsonld_faq.
Input:
Page type: [About Us / Shipping / Returns / Privacy Policy / Size Guide]
Brand summary: [one-sentence]
Contact email/phone: [or "MISSING"]
5) Translate and localize product content (multi-language)
Prompt:
You are a professional e‑commerce translator. Translate these fields from [source language] to [target language] for PrestaShop import: name, short_description_html, description_html, meta_title, meta_description, link_rewrite. Maintain keywords given: [kw1,kw2] and produce culturally-adapted phrasing. Do not change numeric specs, SKUs, codes, units. Output JSON with original and translated fields. If a slug conflicts with language characters, transliterate to Latin characters suitable for URLs.
6) Generate attribute combinations / variations (combinations matrix)
Prompt:
You are a product data specialist. Given attributes and their values, produce all valid combinations and for each combination return: combination_id (incremental), attributes (object), sku, reference, price_adjustment (e.g., +0.00), quantity, image_index (index in provided image list). Use this pattern for SKU: [base_sku]-[abbr_color]-[abbr_size]. If a combination is invalid (e.g., color X not available in size Y), skip it and explain which were skipped at the end in a short list. Output JSON array.
7) Structured data (JSON-LD) for products
Prompt:
Produce a complete JSON-LD Product schema using the supplied product data: name, brand, image URLs, sku, mpn, description (plain text), price, priceCurrency, availability (InStock / OutOfStock), url. Include aggregateRating if rating data provided and offer validReview snippet if reviews provided. Output only JSON-LD code.
Helpful examples (copy/paste)
Example 1 — Single product prompt (quick)
You are a senior e-commerce copywriter. Write short and long product descriptions plus SEO for import. Output JSON:
{
"name":"[Fill]",
"short_description_html":"
Up to 200 chars
",
"description_html":"Overview
...
",
"meta_title":"Max 60 chars",
"meta_description":"Max 155 chars",
"link_rewrite":"product-slug-here",
"image_alts":["alt1","alt2"]
}
Data:
Name: EcoBlend Ceramic Mug
Brand: GreenHome
Primary features: 350ml, dishwasher safe, double-glaze
Target audience: eco-conscious coffee drinkers
Primary keywords: eco mug, ceramic coffee mug
Tone: warm and friendly
Example 2 — Category prompt
Category name: Travel Mugs
Primary keywords: travel mug, insulated mug, coffee to go
Tone: practical, helpful
What to do if the model hallucinates or leaves fields blank
- Add this line to prompt: “If you don’t have explicit data for a required field, put 'MISSING - ask for info' rather than inventing values.”
- Or ask for a verification pass: “List any assumptions you made.”
Extra tips for PrestaShop-specific workflows
- PrestaShop short_description is often shown on category/product list. Make it concise and benefit-oriented.
- Use HTML in description_html (H2, H3, ) but avoid inline ss.
- For link_rewrite, supply a slug already sanitized or instruct the model to create it.
- For images, return a semicolon-separated list of URLs that PrestaShop can fetch or pair each image URL with an alt text.
- When generating CSV for import, keep encoding in UTF-8 and double-quote fields that include commas or HTML.
Testing & Iteration
- Start with 1–3 products, import into a sandbox store, review formatting and character entities (<, >).
- Iterate prompts to adjust tone, length and keyword placement.
- Keep a changelog of prompt versions and outputs for quality control.
If you want, tell me:
- whether you prefer JSON or CSV,
- one example product (or paste 3 rows for bulk),
and I’ll generate a ready-to-import output using one of the templates above.
