Overview

To help you get started quickly, Akeneo provides several ready-made examples of AI Configurations.

These templates illustrate how different use cases can be implemented, from generating SEO content and auditing data quality to maintaining tone consistency across languages or interpreting visual assets.

Each example below includes:

• A short description of the use case
• The key configuration fields
• A sample instruction prompt
• The type of output you can expect

You can copy these examples directly into your AI Configurations page and adapt them to your own attributes, languages, and products.

Example 1: SEO Metadata Generation

Use Case

Automatically create short, optimized SEO meta descriptions based on product attributes such as name, features, and target audience. This is one of the most common enrichment use cases for AI configurations.

Configuration Details

Example Prompt

Role:
Act as an expert SEO copywriter for e-commerce product pages, ensuring factual accuracy and compliance with provided data.

Context:
You will receive product information with the following sources: Product_Type, name, brand, short_description, and supplier. Use only the information provided. Do not invent or infer features. Avoid superlatives or claims unless explicitly supported.

Task:
Write a short, benefit-oriented meta description that naturally integrates relevant keywords derived from the provided data.

Requirements:
- Primary keywords: product type and product name; include brand if present and space allows.
- If key features are sparse or missing, focus on clear core benefits from the short_description and the product type/name.
- If keyword inclusion conflicts with the 155-character limit, prioritise clarity and core benefits first, then primary keywords.

Format:
- Output one single sentence only
- No headings, no bullet points, no quotes, no emojis
- Must end with a clear call to action

Tone:
Use a concise, persuasive, brand-neutral voice that is trustworthy and benefit-led.

Expected Output

Sunglasses for everyday wear that shield your eyes and elevate your look with simple style - discover the right pair for you today.

Example 2: Data Quality Audit

Use Case

Use AI to automatically identify incomplete or inconsistent product data - perfect for spotting missing attributes, logical clashes, or outliers.

Configuration Details

Example Prompt

#Role:
You are a product data quality consultant.

#Task:
Analyze the provided product attributes and list any issues found, including missing values, inconsistencies, or illogical combinations. Suggest an actionable fix for each.

#Output Format:
List each issue in bullet points:
* Attribute: [attribute name]
* Problem: [short description]
* Recommendation: [suggested correction]

Expected Output

- Attribute: Meta Title 
- Problem: Too generic; lacks brand, model, style, gender, or key features. 
- Recommendation: Expand to a descriptive title, e.g., “Unisex Polarized Sunglasses, UV400 Protection, Classic Wayfarer – Black”. 

- Attribute: Meta Description 
- Problem: Nonsensical placeholder text (“FILL THIS IN LATER!”); not SEO-friendly. 
- Recommendation: Replace with a meaningful summary, e.g., “Stylish polarized sunglasses with UV400 protection for glare-free vision and all-day comfort.”  

Example 3: Tone-Controlled Translation

Use Case

Translate product descriptions from one language to another while preserving brand tone and vocabulary.
Ideal for teams who want consistent, high-quality translations aligned with brand voice guidelines.

Configuration Details

Example Prompt

#Role:
You are the Head of Global Content for a luxury fashion brand. Your goal is to translate the text while keeping the tone elegant, aspirational, and refined.

#Tone of Voice:
Sophisticated, descriptive, and brand-aligned. Avoid slang or overly casual phrasing.

#Task:
Translate the following fields from French to English, keeping the same meaning and maintaining the tone described above.

Expected Output

Input:
“Découvrez la pièce maîtresse de notre nouvelle collection prêt-à-porter.”

Output:
"“Discover the centerpiece of our new ready-to-wear collection.”

Example 4: Set Simple or Multi-Select Values Based on Assets

Use Case

Automatically populate color or material attributes based on images or visual references stored in the Asset Manager.
This configuration helps improve accuracy for search filters and product categorization.

Configuration Details

Example Prompt

#Role:
You are an AI-powered visual tagging assistant.

#Context:
Analyze the main product image and identify the single dominant base color of the garment. Ignore small logos or patterns.

#Task:
Return one of the approved color values that exists in my color options.

#Output:
Return only the color name. Do not describe or explain.

Expected Output

Example:
Input: image of a gold sunglasses
Output: Gold

How to Use These Examples

You can use these examples as a starting point for your own configurations:

• Copy one of the example prompts into a new AI Configuration.
• Adjust attribute selections to match your catalog data.
• Test the configuration using the Preview tool.
• Review and iterate the prompt if necessary (or use the Prompt Optimizer).

When satisfied, save and apply the configuration through the Rules Engine.

Best Practices When Using Examples

Start Simple: Begin with a focused task (for example, “Generate SEO descriptions”) before layering complexity.

Adapt to Your Data: Replace attributes and tone with terms familiar to your brand.

Always Test in Preview: Each catalog is unique; verify that the example produces meaningful results for your products.

Iterate with Feedback: Run several sample products, analyze outputs, and refine your prompt wording accordingly.

Next Steps

Once you’ve built or adapted one of these example configurations, explore how to:

• Execute Configurations at Scale using the Rules Engine for automation