How to Identify Product Categories with AI in conbase.ai

Published on July 08, 2025

By Daniel Manco

Introduction

The Goal:
- By the end of this guide, you'll be able to automatically assign accurate categories to thousands of product records in minutes, streamlining your catalog management workflow.
The Problem:
- Manually choosing a category for every product is slow and inconsistent. Teams spend hours sorting items, which leads to mismatches that hurt search, ads, and reporting.
The Solution:
- Conbase.ai's AI pipeline lets you identify product categories at scale with custom prompts that respect your e-commerce taxonomy.
Prerequisites:
- A Conbase.ai account
- A CSV file containing your products (columns like name, description, price, brand)
- Your official category list or taxonomy

Are you here for the first time ? Then watch a full conbase.ai workflow in action

This video walks you through a complete conbase.ai workflow, from input to final output, so you can see exactly how the platform works and how to get the most out of its features.

Step 1: Set Up Your Project and Import Data

Create a New Project

Log into Conbase.ai, click “New Project,” and name it “Product Categorization – 2025-MM-DD.” This keeps everything tidy.

Upload Your Product Data (CSV)

Upload your CSV with up to 10,000 rows. Conbase.ai auto-detects columns.

Make headers clear: product_name, description, brand, price
Strip stray symbols that may confuse parsing
Add columns the AI might need, such as “current_category” if you plan to overwrite only blanks

Step 2: Build Your AI Categorization Pipeline

Add an Action Step

Click “Add Step,” choose “Create New,” and name it “Identify Product Categories.”

Set Up Your Custom Prompt: 3-Part Recipe

Part 1: Instruction

You are an e-commerce taxonomy expert. Assign the most specific category from the provided list to each product. Follow these rules:
- Use the exact category name from the list.
- If the description is vague, choose "Miscellaneous".
- Do not invent new categories.

Allowed categories:
1. Electronics > Audio
2. Electronics > Computers
3. Home & Garden > Kitchen
4. Apparel > Men > Shirts
5. Apparel > Women > Dresses
6. Sports & Outdoors > Fitness
7. Miscellaneous

Part 2: Context

Product Name: {product_name}
Description: {description}
Brand: {brand}
Price: {price}

Part 3: Output

Output Column: ai_category
Output Example: Electronics > Audio

Add at least two examples for best consistency, then click “Save Changes.”

Step 3: Configure and Run the Pipeline

Fine-Tune AI Settings

Model: "GPT-4o"
Processing Mode: "Instant" for testing and for fast execution, "Scheduled" for large batches and 50% less API costs
Integration: “Fill if empty” to avoid overwriting verified categories

Select Rows and Execute

Filter 10 rows for a quick test.
Run the pipeline and review the ai_category column.
If results look solid, select all rows and rerun.

Step 4: Review and Export Your Results

Preview the Generated Categories

Spot-check random items for accuracy.
Ensure categories match your list exactly.
Correct any outliers by editing the prompt or rerunning specific rows.

Export to CSV

Click “Export” in the bottom-right corner.
Download the new CSV with the ai_category column included.
Import back into your store or PIM system.

Conclusion and Next Steps

You’ve automated product categorization, shaving hours off catalog work and improving search relevance.

Time saved: Compare before vs. after categorization speed.
Quality boost: Fewer misclassified items.
Scalability: Same workflow handles future product drops with zero extra setup.

Book Your Free Conbase.ai Demo and see how else you can streamline content operations.