Automated creation of product characteristics attributes using AI
Enter a new era of product performance management
Imagine being able to instantly create a complete set of specifications for any product category—from smartphones to garden furniture. No more spending hours thinking about which attributes are important to buyers. Elbuz's artificial intelligence will do it for you!
Automated attribute creation function using AI — is your personal product expert, available 24/7 and knowledgeable about any product category.
.png)
Why is this feature needed?
Problems it solves:
- Empty product cards — buyers can't find the information they need and turn to competitors
- No filters — customers cannot quickly sort products by the required parameters
- Waste of time — managers spend days manually creating attributes for each category
- Non-systemic approach - different employees create different sets of characteristics, confusion arises
- Lost sales — incomplete product information reduces conversion by 30-50%
- Low SEO — Search engines do not index product cards without structured data
What you will get after using:
- Ready-made set of attributes in 30-60 seconds instead of several hours of work
- Professional approach — AI knows which characteristics are important for each category
- Full product cards — buyers receive all the necessary information
- Convenient search filters — customers quickly find the product they need
- Increasing expertise — detailed specifications make your store reliable in the eyes of buyers
- Conversion growth — complete information increases sales by 25-40%
- Budget savings — automation frees up employees for more important tasks
How does it work? The magic of technology
The system uses advanced artificial intelligence algorithms (GPT-5, Google Gemini) that analyze the category name and automatically generate a list of relevant attributes.
AI work process:
- Category analysis — AI studies the name and path of the category in your catalog
- Checking existing attributes — the system sees which characteristics have already been added
- Generating new attributes — AI creates 25-40 universal characteristics for this category
- Selecting a data type — for each attribute the type (text, number, yes/no) is automatically determined
- Importance assessment — AI indicates how confident it is in the correctness of each attribute selection
- Providing the result - you receive a ready-made list for checking and adjustments
Step-by-step instructions for operation
Step 1: Select a category
Open the category directory. Select the category for which you want to create attributes and click the button. Attribute Search (Creating Category Attributes via AI).
Advice: Start with the categories that have the most products to get the most out of your automation.
Information about the selected category will be displayed on the screen:
- Category name — for example, "Smartphones"
- Category path — full path in the catalog (Electronics → Phones → Smartphones)
- Number of products - how many products are there in this category
Step 2: Start generation
Click the button Create Attributes - this will start the generation process through artificial intelligence.
The following will appear on the screen:
- Progress bar - shows that the system is working
- Operation log — a text log of all AI actions
The wait takes approximately 30-60 seconds depending on the complexity of the category.
Important: Don't close the page while the process is being generated! Wait for the "Operation completed successfully" message.
Step 3: Checking the results
Once generation is complete, the following will appear on the screen:
Statistics panel with 4 cards:
- Total (blue) - total number of attributes created
- Confirmed (green) - attributes ready to be saved
- Waiting (orange) - attributes that require your verification
- Rejected (red) - attributes you have rejected
Attribute table with columns:
- Checkbox — for mass selection of attributes
- Status — a check mark means "confirmed", an empty checkbox means "rejected"
- Attribute name — the name of the characteristic (attributes that already exist in the category are marked in green)
- Type — drop-down list with data types (text, number, yes/no)
- Required — is it necessary to fill in this attribute?
- Confidence - how confident is the AI in the correctness (0-100%)
- Actions — attribute delete button
Clue: Attributes with confidence above 80% (green) generally do not require verification. Note the attributes with low confidence (red and orange).
Step 4: Editing Attributes
You can edit any attribute in the table:
Name change:
- Click on the pencil icon next to the attribute name
- Enter a new name in the modal window
- Click "Save"
Change type:
Select the desired type from the drop-down list in the "Type" column. Changes are saved automatically.
Available attribute types:
- Text — for short text values (e.g. "Color", "Material")
- Number — for numerical values (for example, "Weight, kg", "Power, W")
- Not really — for binary characteristics (e.g. "Waterproof", "Wireless charging")
Confirm/Reject:
Click on the checkbox in the "Status" column:
- Enabled (check mark) — the attribute is confirmed and will be saved
- Off — the attribute is rejected and will not be saved
Mandatory:
Check the box in the "Required" column if this attribute must be filled in for all products in the category.
Removal:
Click the trash can icon in the Actions column to completely remove the attribute from the list.
Step 5: Filtering and Searching
Use the filter bar to easily work with large lists:
Filter by status:
- All — show all attributes
- Awaiting verification - only attributes that require your attention
- Confirmed - only confirmed attributes
- Rejected - only rejected attributes
Filter by type:
Select a specific attribute type (text, number, yes/no) to see only attributes of that type.
Search:
Enter any word in the search field and the system will instantly filter attributes by name.
Convenience: All filters work in real time, results are updated automatically.
Step 6: Bulk Operations
To speed up your work, use bulk operations:
Selecting multiple attributes:
- Check the boxes next to the required attributes in the first column of the table.
- Or click the button Select all to select all attributes on the page
- Button Deselect will uncheck all the boxes
Delete selected:
After selecting the attributes, a button will appear "Delete (N)", where N is the number of selected elements. Click it to delete all selected attributes.
Clearing the entire category:
Button Clear All Clears all generated results for the current category. Use this if you want to start over.
Step 7: Adding Attributes Manually
If the AI didn't suggest an important attribute, you can add it yourself:
- Click the button Add manually
- In the modal window, enter:
- Attribute name — for example, "Equipment"
- Attribute type - select from the drop-down list
- Required - check the box if necessary
- Click "Save"
The new attribute will appear in the table with the status "Confirmed" and 100% confidence.
Step 8: Save to system
Once you have checked and adjusted all the attributes, it is time to save them to the system:
- Make sure that all required attributes have a "Confirmed" status (check mark in the "Status" column)
- Click the button "Save" in the control panel
- Confirm the action in the modal window
- Wait for the message about successful saving
What happens when you save:
- All confirmed attributes are added to the category
- Attributes become available for filling in product cards
- Attributes marked with the "Required" checkbox will automatically appear in the filters on the website.
- The system records that the attributes have been created and do not need to be re-created (check for duplicates)
Important: Only attributes with a "Confirmed" status are saved. Rejected and pending attributes will not be saved.
Step 9: Update and Regenerate
You can at any time:
- Click "Refresh" to reload the attribute list
- Restart generation with the button Create Attributes — AI will add new attributes to existing ones
- Clear the entire category with the button Clear All and generate attributes from scratch
Description of main fields and settings
Category Information
| Field | Description |
|---|---|
| Category name | The name of the category for which attributes are generated (e.g. "Laptops") |
| Category path | The full directory path helps the AI understand the context (e.g., "Electronics → Computers → Laptops") |
| Number of products | How many products are in this category? |
Attribute fields in a table
| Field | Description | How to use |
|---|---|---|
| Selection checkbox | Checkbox for bulk operations | Select multiple attributes for bulk deletion |
| Status | Attribute approved or rejected | Check the box = approved (will be saved), check the box = rejected (will not be saved) |
| Attribute name | Name of the product characteristic | Click the pencil icon to edit. Attributes that already exist in the category are highlighted in green. |
| Type | Attribute data type (text, number, yes/no) | Select the appropriate type from the drop-down list |
| Required | Is the attribute required to be filled in? | Check the box if the feature is critical to the product. |
| Confidence | How confident is the AI in making the right choice (0-100%) | Green (80-100%) = high confidence, orange (60-79%) = medium, red (<60%) = низкая |
| Actions | Attribute control buttons | The trash can icon removes an attribute from the list. |
Filters and search
| Filter | Description | When to use |
|---|---|---|
| Filter by status | Shows only attributes with the selected status. | Use "Pending Review" to view attributes that require your attention. |
| Filter by type | Shows only attributes of the selected type. | Useful for checking all numeric attributes or all checkboxes. |
| Search | Filters attributes by name | Quickly find a specific attribute in a large list |
Control Panel
| Button | Action |
|---|---|
| Create attributes | Runs attribute generation via AI |
| Save | Saves all confirmed attributes to a category |
| Add manually | Opens a form for manually adding an attribute. |
| Refresh | Reloads the attribute list from the server. |
| Delete (N) | Removes the selected attributes (appears when selected) |
| Clear all | Deletes all generation results for a category |
| Select all | Marks all attributes on the current page. |
| Deselect | Unchecks all selection boxes |
Recommendations for use
When to use this feature
- When creating a new category — AI will immediately create a basic set of attributes
- When expanding the range — Added a new product category? Create attributes for it!
- When importing goods - Before bulk import, make sure that all required attributes have been created
- When auditing categories — check if you have everything you need to fully describe the goods
- When improving filters — Do you want to add more filters for the convenience of buyers?
Optimal work strategy
- Check the AI, but trust it — attributes with high confidence (>80%) are usually correct
- Add specifics — AI may not know the specifics of your industry, so add them manually.
- Don't be afraid to experiment - you can always restart the generation
- Save in parts — It is better to keep 10 verified attributes than 50 unverified ones
Tips for choosing attribute types
| Type | When to use | Examples |
|---|---|---|
| Text | For any text values, names, articles | Color, Material, Model, Country of Origin |
| Number | For quantitative characteristics | Weight (kg), Power (W), Diagonal (inches), Volume (l) |
| Not really | For characteristics of the "yes/no" type | Waterproof, Wi-Fi, Bluetooth, Built-in battery |
How to improve the quality of generation
- Give categories clear names "Smartphones" is better than "Goods 123"
- Restart generation - AI may suggest new attributes upon re-launch
- Combine AI and manual input — the best results are achieved by combining automation and expertise
What to do with attributes after saving
- Fill in the values - open product cards and fill in new attributes
- Set up filters - Check that the attributes are displayed in the filters on the site
- Train the team — tell employees about the new attributes and why they are needed
- Monitor occupancy - Make sure that attributes are filled in for all new products
- Analyze usage — see statistics on what filters buyers use
Frequently Asked Questions
Is it possible to edit attributes after saving?
Yes, once saved, attributes become regular system attributes. You can edit their names, types, and values in the attribute management section.
What if AI created inappropriate attributes?
Simply uncheck the "Confirmed" status for any attributes you don't need or delete them. They won't be added to the category when you save them.
Is it possible to start generation for all categories at once?
Yes, there is a feature for bulk attribute creation for all categories. However, it is recommended to use it only for initial populating, as the results are not manually verified.
How many attributes does AI create at a time?
The AI generates between 25 and 40 attributes depending on the category. Complex categories (such as electronics) receive more attributes, while simple categories (such as accessories) receive fewer.
Does AI take into account existing attributes?
Yes! The system analyzes the attributes already present in the category and creates only new ones. Existing attributes are shown in green in the list.
What does percentage confidence mean?
This is the AI's assessment of how appropriate a given attribute is for a category. The higher the percentage, the more confident the AI is. Attributes with a confidence score below 60% are recommended for particularly careful review.
Is it possible to change the number of attributes generated?
In the current version, the number of attributes is fixed (25-40). However, you can run the generation multiple times—the AI will add new attributes to the existing ones.
Are the results saved if I close the page?
Yes, all generated attributes are saved in the system. You can close the page and return to the check later.
What should I do if generation fails?
Check your internet connection and try again. If the error persists, contact technical support—there may be a temporary issue with the AI service.
Is it possible to roll back the saving of attributes?
There is no automatic rollback, but you can manually remove unwanted attributes from a category in the attribute management section.
Examples of use
Example 1: Category "Smartphones"
Task: Create attributes for the smartphone category.
What AI did:
- Screen diagonal, inches (number, 95% confidence)
- RAM capacity, GB (number, 98% confidence)
- Built-in memory capacity, GB (number, 98% confidence)
- Number of main cameras (number, 90% confidence)
- Main camera resolution, MP (number, 92% confidence)
- Battery capacity, mAh (number, 95% confidence)
- Operating system (text, 88% confidence)
- Case material (text, 85% confidence)
- 5G support (yes/no, 93% confidence)
- Waterproof (yes/no, 90% confidence)
- Wireless charging (yes/no, 88% confidence)
- ...and 25 more attributes
Operator actions:
- I checked all the attributes - everything is correct.
- Manually added the "Complete set" attribute (text)
- Marked "Screen Diagonal" and "Operating System" as required
- Clicked "Save"
Result: 37 attributes were saved to a category in 5 minutes instead of 2-3 hours of manual work.
Example 2: Category "Outdoor Furniture"
Task: Create attributes for a new category of garden furniture.
What AI did:
- Frame material (text, 92% confidence)
- Upholstery material (text, 88% confidence)
- Maximum load, kg (number, 90% confidence)
- Number of seats (number, 85% confidence)
- Weight, kg (number, 87% confidence)
- Folding (yes/no, 93% confidence)
- Water-resistant upholstery (yes/no, 91% confidence)
- UV resistance (yes/no, 89% confidence)
- Backrest adjustment (yes/no, 86% confidence)
- ...and 20 more attributes
Operator actions:
- I checked the attributes - everything is fine.
- Added specific attributes: "Frame Color", "Upholstery Color", "Style"
- Rejected the "Backrest Adjustment" attribute - not relevant for tables
- Clicked "Save"
Result: The new category has 32 quality attributes, allowing buyers to easily filter products.
Example 3: Category "Sports Nutrition"
Task: Supplement existing attributes with new ones.
What was in the category:
- Package weight, g
- Taste
- Manufacturer
What AI added:
- Protein content per serving, g (number, 96% confidence)
- Carbohydrate content per serving, g (number, 94% confidence)
- Fat content per serving, g (number, 94% confidence)
- Calories per serving, kcal (number, 95% confidence)
- Number of servings per package (number, 93% confidence)
- Release form (text, 88% confidence)
- Lactose-free (yes/no, 90% confidence)
- Gluten-free (yes/no, 89% confidence)
- Vegan (yes/no, 87% confidence)
- ...and 18 more attributes
Operator actions:
- I saw existing attributes marked in green - excellent, no duplicates
- Confirmed all nutritional characteristics - this is critically important
- Manually added "Certification" (text)
- Marked food characteristics as mandatory
- Clicked "Save"
Result: The category received a full set of nutritional characteristics, which is important for sports nutrition buyers.
Tips for optimal configuration
For maximum efficiency
- Standardize names
- Use a single format: "Characteristic, unit of measurement"
- Examples: "Weight, kg", "Power, W", "Volume, l"
- This will make it easier to import and export data.
- Group attributes by meaning
- Technical specifications (dimensions, weight, power)
- Functionality (what the product can do)
- Materials and appearance (what it is made of, what it looks like)
- Equipment (what is included in the set)
- Minimize required attributes
- Mark only critical features as required
- Too many required fields slow down adding products
- Typically 3-5 required attributes per category are sufficient.
- Conduct audits regularly
- Once a quarter, run the generation again to update
- AI can suggest new relevant attributes
- Remove unused attributes to keep your system clean
For different types of business
Online electronics store
- Focus on technical characteristics (numeric attributes)
- Add as many attributes as possible - electronics buyers love details
- Use required attributes for key parameters
Clothing and footwear store
- Focus on sizes, colors, materials (text attributes)
- Add attributes for filters: season, gender, style, purpose
- Минимум обязательных атрибутов — главное размер и цвет
Grocery store
- Обязательно включайте пищевую ценность (числовые атрибуты)
- Добавляйте сертификацию, состав, срок годности
- Используйте чекбоксы для диет: без глютена, веганское, халяль и т.д.
Магазин мебели и товаров для дома
- Акцент на размерах и материалах
- Добавляйте атрибуты стиля и цветовой гаммы
- Указывайте особенности ухода и эксплуатации
B2B supplies
- Максимальная детализация технических параметров
- Добавляйте артикулы, коды, стандарты, сертификаты
- Указывайте параметры упаковки и логистики
Заключение: Ваш путь к идеальному каталогу
Автоматизированное создание атрибутов с помощью AI — это не просто инструмент. Это ваш шаг в будущее e-commerce, где технологии работают на вас, освобождая время для стратегических задач.
Что вы получаете с этой функцией:
- Экономию 90% времени на создании атрибутов
- Professional quality описания товаров
- Счастливых покупателей, которые легко находят нужный товар
- Sales growth благодаря полной информации
- Конкурентное преимущество at the market
Start now!
- Выберите категорию с наибольшим количеством товаров
- Нажмите "Создать атрибуты"
- Проверьте результат за 2-3 минуты
- Сохраните и наслаждайтесь результатом
Remember: Искусственный интеллект — это ваш помощник, но финальное решение всегда за вами. Комбинируя мощь AI и вашу экспертизу, вы создадите идеальный каталог, который приведёт к росту продаж и лояльности клиентов.
"Лучшее время, чтобы начать использовать AI, было вчера. Второе лучшее время — сегодня."
Успехов в автоматизации вашего интернет-магазина! 🚀
Техническая поддержка и обратная связь
Если у вас возникли вопросы или предложения по улучшению функции, свяжитесь с технической поддержкой Elbuz.
Мы постоянно совершенствуем наши AI-инструменты и будем рады услышать ваши отзывы!
- Enter a new era of product performance management
- Why is this feature needed?
- How does it work? The magic of technology
- Step-by-step instructions for operation
- Description of main fields and settings
- Recommendations for use
- Frequently Asked Questions
- Examples of use
- Tips for optimal configuration
- Заключение: Ваш путь к идеальному каталогу
- Техническая поддержка и обратная связь

