Hi,
I’m trying to figure out the best design/diagram/suggestion for a workflow, I guess with an AI agent, to handle the following tasks.
The Context
- Business : Sells thousands of spare parts online via Shopify
- Suppliers: Dozens
- ERP : Uses Zoho Inventory to manage stock, pricing, and updates
Current Challenge
- Supplier price updates are sent as Excel files with multiple tabs (one per product category)
- Each tab contains (at minimum) two key columns:
- General Item No
- Purchase price
- Manual entry of these values into Zoho Inventory is slow and error-prone
Target Workflow
- User triggers the workflow manually (e.g., by clicking an icon)
- A chat-like interface welcomes the user and prompts for Excel upload
→ (Note: This should be handled outside of n8n, since the client doesn’t have access to n8n UI. Use a simple web interface, Telegram bot, or similar.) - Once uploaded:
- Read all Excel tabs
- Extract only General Item No and Purchase price columns from each tab
- Create a new summary tab consolidating these two columns from all other tabs
- Perform data validation:
- Highlight rows with issues in orange (e.g., empty values, wrong formats)
- Generate a “Data to be Checked” report with:
- Tab name
- General Item No
- Issue description
- Present summary back to user via chat
- Wait for user to correct the Excel file , then type update
- Reload the file and reprocess clean data
- Call Zoho Inventory API to:
- Update Purchase Price for each item
- Update Selling Price (basic formula)
- Notify user with: “Your Zoho database has been successfully updated.”
Additional Notes
- Zoho API access is already authorized
- File upload location is TBD: open to recommendations between local processing vs. using Google Drive, etc.
- Chat interface should be external to n8n (not the default n8n frontend, probably Telegram)