I have two Excel files
- File A: contains existing SOP names already in the system
- File B: contains new SOP names that will be uploaded
The names may not match exactly, but they often mean the same thing. For example, “Monthly financial report summary” vs “Prepare monthly finance report”.
I want to build a workflow in n8n that:
- Reads both Excel files
- Compares each new SOP name with existing ones
- Detects if any are semantically similar (even if wording is different)
- Returns possible matches or duplicates
What is the best approach to achieve this in n8n? And how should I handle performance if I have hundreds of SOP entries?
Any guidance or example workflows would be appreciated.
1 Like
If you’re trying to detect whether any new SOPs are similar to existing ones (even if the names aren’t exactly the same), you can actually use AI-powered semantic matching in your workflow. Use an AI Agent or Text Similarity Tool
OpenAI embeddings (with cosine similarity)
Hugging Face sentence transformers
Cohere or other semantic similarity APIs
These help catch things like:
“Monthly financial report summary” ≈ “Prepare monthly finance report”
A Basic Workflow could look like this:
Read both Excel files (new and existing SOPs).
Loop through each new SOP.
Compare it against the existing list using an AI text similarity tool.
Return matches above a certain similarity threshold (e.g., 85%).
You could also use a lightweight AI agent in n8n with a prompt like:
“Does this SOP already exist in the following list? Return any similar ones.”
Hope this helps.