How to Configure n8n to Process Only New Updates and Align AgentsAI Responses on the Same Line?

Hello! How are you? I would like some help with the tool n8n. Overall, my workflow works, but I’m encountering some problems and spinning in circles without finding a solution. I’ll explain the workflow to provide some context:

The goals of my workflow are:

1. Data Collection: I use Google Forms to automatically generate a Google Sheets spreadsheet with my clients’ responses.
2. Reading: I retrieve the responses from the spreadsheet through n8n.
3. Preparation: I unify the new updates in the spreadsheet (transforming all the questionnaire responses into a single text) so that my multiple AI agents can perform an analysis (with “agents” in the plural, as the idea is to have several AI agents, each providing a different and specific analysis based on the responses).
4. Analysis: Each AI agent executes a customized prompt that defines its personality, functions, objectives, and expected results.
5. Result Generation: At the end of the process, a new spreadsheet is created, containing the output from each AI agent, with each column representing the data processed and analyzed individually.
.

Problems I’m facing with the final result:

Problem 01: How to Process Only New Updates to the Spreadsheet?
I need the workflow to process only the new updates in the spreadsheet, avoiding the re-analysis of already processed responses. Currently, whenever the spreadsheet is updated, the workflow reprocesses all responses, including those previously analyzed. I need the workflow to ignore already processed responses and focus solely on the new entries added after the last execution.

Problem 02: How to align Agent Responses on the Same Line?
AI agent responses are being stored in different rows on the final spreadsheet, causing misalignment and confusion.
For example, after a form update, Agent 01’s response goes to row 2, Agent 02’s response to row 3, Agent 03’s response to row 4, and so on. In the next entry, Agent 01’s response will be placed in row 5 (which is an issue).
Ideally, all responses from agents for the same entry should be on the same row, but in different columns. This means each row in the spreadsheet should represent a single form entry, with each column showing the output from a different AI agent.

It looks like your topic is missing some important information. Could you provide the following if applicable.

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:

For Problem #1:

You could add a binary column to your sheet which you can update after successful processing and also filter by when grabbing new entries. Something like “processed”: true|false

For Problem #2:

I would highly recommend moving from Google Sheets to a relational database to ensure data consistency. There you can build solid structures and address records by a unique ID which should solve your problem. You could use Supabase, MySQL, Airtable, NocoDB, just to name a few.

1 Like

Thanks for answering me and for your tip! I’m trying to implement it, but I’m facing some difficulties.

I created a new column in my spreadsheet called “statusstamp” to identify whether or not a row has been previously parsed. I was able to filter this information with N8N, but I’m having trouble updating the spreadsheet with the “true” information in the “statusstamp” column.

Could you help me resolve this?

Thank you in advance for your help!

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.