Build an AI Product Data Enrichment Automation with n8n, Airtable, and OpenAI GPT

Introduction & What This Automation Does

In this project, we build a complete AI Product Data Enrichment Automation using n8n, Airtable, and OpenAI. The goal of this automation is to transform raw product data into rich, structured, and high-quality product content automatically using AI.

This type of automation is extremely useful for e-commerce platforms, marketplaces, internal product catalogs, and product management systems that require consistent and detailed product information. Instead of manually writing descriptions, SEO titles, summaries, and marketing content, the entire process is handled automatically by the workflow.

In this system, Airtable acts as the main database where all product data is stored. Airtable serves both as the input source for raw product data and the destination where the enriched product content is written after processing.

You can view and copy the Airtable base used in this project using the public link below. This base contains the same Product table structure and fields used in the automation, so you can easily duplicate it and use it with the n8n workflow without needing to recreate the database schema manually. Simply open the link, copy the base to your Airtable workspace, and you will have the exact structure required to run this automation.

https://airtable.com/app5664nhfEFhTGT9/shr0Fx6RlTIzV3FEz

With just four input fields — product name, product category, brand name, and original description — the workflow automatically generates:

• SEO title

• Full product description

• Bullet points

• Short description

• Keywords

• Value proposition

• Benefit list

• Product summary

• Use cases

• Target audience

All generated data is automatically written back to Airtable, creating a fully enriched product record.

If you want to build this automation yourself, you can download the complete n8n workflow JSON from the link below and import it directly into your n8n instance.

https://drive.google.com/drive/folders/1xBovUqo_QXrV17ARwVYH41xJg-GtSq0S?usp=drive_link

Quick Demo ⇒ Youtube link embedded

You can watch the full quick demo of this project below.

https://www.youtube.com/watch?v=S4Pots2Pc7k

Technologies Used and Prerequisites

This project uses several tools that work together to automate the product enrichment process.

n8n is used as the automation engine that connects all services and handles the workflow logic.

Airtable is used as the central database where all product data is stored and updated.

OpenAI is used to generate the enriched product content using a structured prompt and LLM chain.

Below are the main technologies used in this project:

• n8n

• Airtable

• OpenAI API

• Structured Output Parser

• n8n Form Trigger

Before using this workflow, make sure you have the following:

• An n8n instance (self-hosted or n8n cloud)

• An Airtable base with the required product fields

• An OpenAI API key

• Basic understanding of how to import workflows into n8n

Workflow Explanation

This automation consists of a main workflow with two triggers, allowing the system to process products in different ways depending on the use case.

Manual Trigger Flow (Batch Processing)

The first trigger is a manual trigger that processes all products with a pending status.

The workflow begins with the Search Products with Pending Status node. This node retrieves all products from Airtable that are marked as pending. This ensures that only products that have not yet been processed will be enriched.

Next is the Limit node. This node is currently disabled, but it allows you to limit the number of products processed in a single execution. This is extremely useful when working with large product catalogs, since you may want to process items in batches, such as 10 products per run.

After that, the workflow enters a Loop node, which processes each product individually. This ensures that every product is handled correctly and that the AI generation process runs independently for each record.

AI Content Generation

The core of the automation is the Generate Enriched Product Content LLM Chain node.

This node sends product information to the AI model and generates structured product content. The prompt uses the following inputs:

• Product Name

• Product Category

• Brand Name

• Original Description

The AI returns a structured JSON object containing all enriched product fields such as descriptions, bullet points, summaries, and marketing content.

The workflow uses OpenAI GPT-4.1 Mini, but you can easily switch to other models depending on cost, speed, or quality requirements.

To ensure the output remains consistent, the workflow uses a Structured Output Parser. The parser enforces the same JSON structure that is defined inside the prompt. This ensures the generated content always follows the expected format, which is critical for updating Airtable correctly.

Updating the Airtable Database

After the AI content is generated, the workflow moves to the Update Product in Airtable node.

This node writes all generated content back into the product record inside Airtable and updates the product status to done.

Some fields use custom expressions for formatting. For example, the Bullet Points field converts an array into a numbered list so the output is easier to read inside Airtable. Similar formatting logic is used for:

• Keywords

• Value propositions

• Benefit lists

• Target audience

This ensures the final data is both structured and visually clean.

n8n Form Trigger Flow

The second trigger uses an n8n Form. This trigger allows users to process a specific product by entering its Product ID.

The form contains a single required field:

• Product ID

After submission, the workflow runs the Search Product by Product ID node, which retrieves the exact product from Airtable.

The workflow then connects to the same enrichment and update logic used in the manual trigger flow. Since Product ID is unique, this process only enriches a single product.

Results in Airtable

After the workflow finishes processing, all enriched content appears directly in the Airtable product table. Each product record now contains complete product information including descriptions, summaries, marketing content, and SEO fields.

The product status is also updated to indicate that the enrichment process has been completed.

Summary

This project demonstrates how AI and workflow automation can be combined to automatically enrich product data at scale.

Instead of manually writing product descriptions and marketing content, the system generates everything automatically from a small set of input fields.

The automation includes:

• A manual trigger for batch processing products with pending status

• An n8n form trigger for processing a specific product on demand

• An AI enrichment pipeline powered by OpenAI

• A structured output parser to guarantee consistent data formatting

• Automatic Airtable updates for storing the enriched content

With this setup, businesses can dramatically improve the quality and consistency of their product data while saving significant time on manual content creation.

This approach can be easily adapted for large e-commerce catalogs, marketplaces, and internal product databases, making it a powerful example of how AI automation can streamline real business workflows.

If you want to build this automation yourself, you can download the full n8n workflow JSON from the video description and import it directly into your own n8n instance.

Thanks for reading, and I'll see you in the next project.