Automate Product Data Enrichment with n8n and AI Tools

What This Automation Does

This workflow scans a list of building product photos stored in Airtable that need more information.

It uses OpenAI vision model to look at pictures and finds product details like description, model number, what it is made of, color, and how worn it is.

Then, an AI agent tries to get more product facts by searching images online via SERP API and scraping product pages with Firecrawl.

All this gathered data is cleaned up and sent back to Airtable to update the product records automatically.

This helps save lots of manual work and improves the data correctness without the user needing to do extra research.

Who Should Use This Workflow

This workflow fits best for people who survey buildings and want to quickly fill product info from photos.

Users who keep product photos in Airtable and want to cut down hours of manual typing will find this useful.

The workflow suits those wanting smarter and faster data enrichment that is guided by AI with online searching support.

Tools and Services Used

• Airtable API: Reads and updates product photo rows.
• OpenAI Vision Model (GPT-4o vision): Analyzes image content for product info.
• OpenAI ChatGPT model with AI Agent: Controls logic for deeper research and uses tools.
• SERP API Google Reverse Image Search: Finds similar product images on the internet.
• Firecrawl Scrape API: Extracts relevant text from product webpages in markdown format.
• n8n workflow automation platform: Coordinates all API calls and data updates.

Beginner step-by-step: How to Use in n8n Production

Step 1: Import the Workflow

1. Download the workflow file using the Download button on this page.
2. Open your n8n editor where your automations live.
3. Click on the menu and choose “Import from File.”
4. Select the downloaded workflow file to load it into n8n.

Step 2: Add Credentials and Configure

1. Update the Airtable Base ID and Table ID in the Get Applicable Rows node.
2. Add your API Keys for OpenAI, SerpAPI, and Firecrawl in the Credentials section.
3. Check other fields like email, channel, or folder IDs if applicable and update them.

Step 3: Test the Workflow

1. Click the manual trigger node called Webhook node or Manual Trigger.
2. Run the workflow and watch logs for each step.
3. Verify Airtable rows get updated with enriched product data.

Step 4: Activate for Production

1. Once testing passes, turn on the workflow in n8n so it runs on schedule or from webhook triggers.
2. Monitor executions in the n8n dashboard and fix any errors.

Use this simple import then configure method without needing to build from scratch.

For users running self-host n8n, check self-host n8n for hosting options.

Inputs, Processing, and Outputs Explained

Inputs

• Airtable rows with product photos and AI_status=false.
• Photos URL from Airtable image fields.

Processing Steps

• OpenAI Vision Model analyzes photos for basic product info.
• AI Agent checks if more data needed and calls SERP API reverse image search.
• Agent uses Firecrawl to scrape product page content found online.
• Responses from APIs parsed and structured into product data attributes.
• Routing logic controls which API is called based on AI agent tool use.
• Error handling nodes cover failures or API unavailability.

Outputs

• Airtable rows updated with enriched fields: description, model, material, color, condition.
• Boolean flag AI_status set to true to mark processed records.

Customization Ideas

• Change the image analysis prompt in the Analyse Image node to get more or different product attributes.
• Add new research tools into the AI agent by connecting new tool workflows.
• Adjust Airtable filtering to process only certain categories or dates.
• Switch Airtable Base or Table IDs to target different data sets.
• Tailor AI agent prompt system messages to guide its research or data formatting.

Handling Errors and Edge Cases

Edge case: No rows fetched from Airtable

Check Base and Table IDs are correct and API access is live.

Make sure filter formula includes records with images and AI_status false.

Edge case: OpenAI Image Analysis fails

Confirm image URL points to large thumbnails and is correctly referenced.

Check OpenAI API Key is set in credentials and not expired.

Edge case: SERP API rejects requests

Verify API key and query params for reverse image search node.

Use the correct custom HTTP request node, do not use generic SERP nodes without params.

Edge case: Firecrawl Scrape API times out or errors

Check Content-Type header is ‘application/json’.

Review API key and request payload format carefully.

Watch for network or firewall blocking on server.

Pre-Production Tips and Deployment

• Test all API credentials separately before full run.
• Run workflow on a single sample record with a valid photo first.
• Backup Airtable data before bulk updating.
• Activate workflow for scheduled or webhook triggers after testing.
• Monitor logs and performance within the n8n interface.

If using self-host n8n environment, add this resource: self-host n8n.

Summary of Benefits and Outcomes

✓ Saves 8-12 hours per week on manual data entry.
✓ Improves accuracy of product information.
✓ Automatically updates Airtable without user typing.
✓ Uses AI vision plus online search for complete data.
✓ Handles hundreds of records effectively.
✓ Reduces human errors and project delays.

→ Gives building surveyors a faster way to enrich product data from photos.
→ Frees time to focus on analysis or client work.
→ Provides a simple but smart automation powered by OpenAI and trusted APIs.