I’m working on a proofreading workflow in n8n and I’ve run into a small but super frustrating problem.
Here’s what I’m trying to do:
Pull data from a Google Sheet using SplitInBatches.
Each row has just one word in it.
I send that word to the Groq Chat model (basically GPT-4) to check if it’s correct.
If the word is correct → update the sheet with “Ignore” in column G.
If it’s wrong → update with “Mark”.
Everything works fine except the last step.
The AI correctly figures out if the word is right or wrong, and I get the expected result (“Ignore” or “Mark”). But for some reason, it doesn’t update the sheet. No errors, the workflow runs successfully, but nothing changes in the actual sheet.
Things I’ve checked:
Column G exists and has dropdown values “Ignore” and “Mark”.
The bot has edit access to the sheet.
I logged the AI output before updating – it looks correct.
Just can’t figure out why the value isn’t going into the sheet
Anyone faced something like this before? Would love any tips or pointers. Thanks in advance!
Hey @cybernedis ,
What n8n version are you on and waht is the environment (cloud, Community on Docker/npm)? Asking because I do not recognize the SplitInBatches node.
Normally you want Loop Over Items to be an entry point to the loop, meaning it goes right after input data formation (SplitIntoBatches or Edit Fields). Example: Loop Over Items (Split in Batches) | n8n Docs
In your current scenario AI Agent would be able to process only first item from its input and Loop item will start processing items that come from the second Sheet node.
There might be other issues in your workflow, but the above should be fixed first.
Try to fix it. If this doesn’t help, then pin output data in readTool and AI Agent node and post the workflow using </> button.
i’m on 1.85.4, community build on docker, self hosted. I have like fixed that by adding an edit field after the ai agent, then also added if after the loop, it worked fine , now the issue i’m facing here is this. I’m a total beginner so playing around
