Below is a practical, copy‑paste‑ready guide you can use to write “perfect” AI prompts for PrestaShop product pages, category pages, and CMS pages. It gives a clear prompt structure, ready templates (short/long, SEO, JSON/CSV export), sample filled prompts, and tips for iteration and quality control.
Why this structure works
- Be explicit about role, goal, inputs, outputs, constraints, and s.
- Give the AI required fields and optional fields and a strict output format (JSON/CSV/HTML) so results are import-ready.
- Ask for multiple variants (A/B) and lengths (short/long) so you can test what converts best.
- Tell the AI what it must not invent (e.g., technical specs) and what to do when information is missing.
Core prompt structure (use for every task)
1. Role: “You are an expert e-commerce copywriter / SEO specialist / PrestaShop content manager.”
2. Goal: one-sentence objective (e.g., “Create product descriptions and meta tags for PrestaShop import.”)
3. Inputs: explicit list of fields I will supply (product name, features, specs, target audience, keywords, brand voice, language, max lengths).
4. Outputs: exact fields + format (JSON, CSV, HTML). State required keys/columns used by PrestaShop.
5. Style & voice: tone, reading level, special instructions (e.g., include X, avoid Y).
6. SEO constraints: primary keyword, secondary keywords, meta lengths, unique meta title and description.
7. Safety/accuracy rule: do not invent specs; if info is missing, say “MISSING:
” or ask a clarifying question.
8. Variation requests: e.g., “Provide 3 short descriptions, 2 long descriptions, 2 meta titles, and 2 meta descriptions.”
9. Examples: show a short example of expected output for one item.
Product description prompt templates
A. Import-ready JSON template (single product)
Use when you want a single product output in JSON to import or map to PrestaShop columns.
Prompt:
You are an expert e-commerce copywriter and SEO specialist. Create import-ready JSON for a PrestaShop product using the input fields below. Do NOT invent any technical specs or dimensions—if a field is missing output null for that field. Produce only valid JSON (no extra commentary). Use the language: {language}. Use the brand voice: {brand_voice}.
Inputs:
- name: {product_name}
- brand: {brand}
- short_features: {short_bullets} (3–6 short bullet points)
- technical_specs: {specs} (key:value pairs)
- product_type/category: {category}
- primary_keyword: {primary_keyword}
- secondary_keywords: {comma_separated}
- target_audience: {target_audience}
- color/variant_info: {variants}
- images: {image_urls}
- price: {price}
- SKU: {sku}
- shipping_info: {shipping}
- other_notes: {notes}
- max_meta_title_chars: 60
- max_meta_desc_chars: 155
Output JSON fields required:
- name
- link_rewrite (SEO friendly URL slug)
- meta_title
- meta_description
- description_short (HTML, max 300 chars)
- description (HTML long)
- tags (comma separated)
- features (array of {name, value})
- attributes (array for size/color if provided)
- images (array of URLs)
- price
- sku
- reference (same as sku)
- quantity (if provided else null)
- id_category_default
- additional_categories (array)
- active (true/false)
- visibility (both/catalog/search)
- shipping (string)
- manufacturer (brand)
Produce:
1) JSON object for the product.
2) Three variations of description_short (25–45 words) and two variations of description (150–300 words) inside the JSON.
If a field is missing, set it to null.
Example (replace placeholders with real values):
{ "name": "EcoBrew Stainless Travel Mug 500ml", "link_rewrite": "ecobrew-stainless-travel-mug-500ml", ... }
B. CSV-ready prompt for bulk import (multiple products)
Use when you have a table and want AI to output CSV rows.
Prompt:
You are an e‑commerce copywriter. Given this CSV header line (do not add extra columns):
name,link_rewrite,meta_title,meta_description,description_short,description,tags,features,images,price,sku,reference,quantity,id_category_default,active,visibility,manufacturer,shipping
For each product below (products will be provided one per line as: name|brand|short_features|specs|category|primary_keyword|price|sku|images|quantity), produce one CSV row matching the header. Fields containing commas should be quoted. Keep meta_title <= 60 chars and meta_description <= 155 chars. description_short <= 300 chars. For features convert short_features into semicolon-separated pairs (FeatureName:Value). Generate link_rewrite from name (lowercase, hyphenated). If any spec is missing, leave value empty. Output only CSV rows (no header).
Category page prompt template
Prompt:
You are an SEO content writer specializing in e-commerce category pages. For the category: {category_name} (parent category: {parent_category}) and target keywords: {primary_keyword}, {secondary_keywords}, generate the following in {language}:
- Meta title (<= 60 chars)
- Meta description (<= 155 chars)
- Category page H1 (one line)
- Intro paragraph (50–80 words)
- 3 short selling bullet points (each 8–12 words)
- Two short SEO-friendly product teasers (each 25–40 words) suitable for listing previews
- Suggested category-specific tags (comma separated)
- Suggested internal link anchor texts to other categories or CMS pages (3 links)
Constraints: Use the primary keyword once in the meta title and once in the first sentence of the intro paragraph. Keep tone: {brand_voice}. Do not invent product-specific specs.
CMS pages prompt template (About, Shipping, Returns, Terms, FAQ)
Prompt:
You are a professional content writer specialized in e-commerce CMS pages. Write a {page_type} page for brand {brand} in {language}. Include:
- A clear H1
- Short intro (40–80 words)
- Sections with H2 headers for each important topic (use 3–6 sections total)
- One FAQ block with 5 Q&A items (each answer 30–60 words)
- A call-to-action at the end (e.g., contact/email/phone)
- Meta title and meta description (<=60/155 chars)
- Internal links suggestions (3 internal links with anchor text)
Tone: {brand_voice}. Include trust signals (warranty, secure checkout, customer support) and shipping/returns timelines if provided. If not provided do not assume.
Examples (filled-in samples)
1) Product example
Input (for the single-product JSON template):
- language: English
- brand_voice: Friendly but professional
- product_name: EcoBrew Stainless Travel Mug 500ml
- brand: EcoBrew
- short_bullets: Keeps drinks hot up to 6 hours; Leak‑proof lid; Dishwasher safe
- specs: Capacity:500ml; Material:18/8 stainless steel; Insulation:Double-wall vacuum
- category: Travel Mugs
- primary_keyword: stainless travel mug
- secondary_keywords: insulated travel mug, leak-proof mug, 500ml mug
- target_audience: commuters, travellers, eco-conscious buyers
- price: 24.99
- sku: EB-TM-500
- images: https://example.com/img1.jpg,https://example.com/img2.jpg
- quantity: 120
- id_category_default: 45
- shipping: Free shipping over $50; standard shipping 3–5 business days
Example instruction (actual prompt):
[Use the JSON template prompt above and paste these inputs.]
Expected type of JSON output (illustrative excerpt):
- meta_title: "EcoBrew Stainless Travel Mug 500ml | Insulated & Leak‑Proof"
- meta_description: "Keep drinks hot up to 6 hours with the EcoBrew 500ml stainless travel mug. Leak‑proof lid, dishwasher safe. Perfect for commuters."
- description_short: "EcoBrew 500ml insulated travel mug — keeps drinks hot up to 6 hours. Leak‑proof and dishwasher safe.
"
- description: longer HTML with features, care, usage, warranty info, CTA.
2) Category example
Input:
- category_name: Travel Mugs
- parent_category: Drinkware
- primary_keyword: travel mugs
- secondary_keywords: insulated travel mugs, stainless travel mug
- brand_voice: Casual, trustworthy
- language: English
Expected outputs: meta title, meta desc, H1, intro paragraph, bullets, product teasers, tags, internal links.
Practical tips and best practices
- Provide as much accurate structured input as possible (specs, materials, warranty). AI must not invent specs.
- Use exact character limits for meta fields and short descriptions.
- Ask for both short (teaser) and long (detailed) variants for A/B testing.
- For multilingual stores: ask the AI to translate and localize (date formats, units). Example: “Translate and localize to French — convert imperial to metric.”
- For consistency: feed the AI a brand s sheet (voice, banned words, preferred phrases).
- For bulk: break your catalog into batches of 50–200 items per request to keep outputs manageable.
- Validate titles/URLs: ensure link_rewrite is unique. You can have AI add a fallback suffix if duplicate detected (e.g., -1, -2).
- Add a line to the prompt telling the model to produce machine-parseable output only (JSON/CSV) to avoid extra commentary.
Quality control steps after generation
- Check for hallucinations: verify all specs against product data sheet.
- Run SEO checks: length, keyword usage, uniqueness.
- Validate HTML: ensure valid tags and no inline scripts.
- Test import on a staging PrestaShop before production.
- Review tone and compliance (legal claims, safety).
Short ready-to-use prompts (copy-paste)
A. Short product description + SEO (one-off)
You are an expert e-commerce copywriter and SEO specialist. Create:
1) A meta title (<= 60 chars) and meta description (<= 155 chars) using the primary keyword: {primary_keyword}
2) One short product description (1–2 sentences, 20–35 words)
3) One long product description (120–220 words) in HTML with 3 bullet points (ul/li). Use brand voice: {brand_voice}. Do not invent specs; if missing, write: "Specification missing: {spec_field}". Product data: name: {name}; features: {short_features}; specs: {specs}; images: {images}.
B. Bulk CSV instruction (for multiple items)
You are an e-commerce writer. Given header: name,link_rewrite,meta_title,meta_description,description_short,description,tags,features,images,price,sku,quantity,id_category_default,active,visibility,manufacturer,shipping
For each product line I provide as: name|brand|short_features|specs|category|primary_keyword|price|sku|images|quantity, produce a CSV row matching the header. Keep meta limits: 60/155. Output only CSV rows.
Closing checklist (before running)
- Have exact keywords and brand voice.
- Decide languages and localization rules.
- Prepare CSV/JSON template you will import into PrestaShop.
- Test on a small batch.
- Audit outputs for accuracy before publishing.
If you want, paste one product (or a small list of products) here and I’ll generate a ready-to-import JSON or CSV row using the template most convenient for your workflow (single product JSON, or CSV for bulk import), and produce meta/title/short/long descriptions in your desired tone and language.
