Describe the problem/error/question
I would like to automate this process of generating action plans as per email from customers. I am tyring to figure out how to do that. That would be appreciated if you point me to the similar threads or effective ways to implement what I would like to do.
Please share your workflow
I am trying to automate the following steps
1. From Slack, I would like to ask n8n to trigger a gmail node to explore my gmail inbox messages that have piled for the last 30 minutes
2. Summarize the context of emails with Field Sets
3. Classify email into relevant topics using Text Classifier
4. Add a label to email based on the classifier
5. Summarize action plans based on the classifer. For example, if an email is from potential customers, then the next node, OpenAI node, summarizes what needs to be done based on the prompt
6. Send the action plans to the team member via Slack if the GPT node finishes summarizing
7. This team member responds to this messege sent from n8n on Slack
You can start from this trigger
It will detect if there are new email coming every 30 minutes.
Then doing some if or filter to check it’s from the customers.
After that doing the thing you need.
The fun of automation is you can create on your own.
But if you really stuck on the next step. I can give you the approximate workflow then.
@darrell_tw
Thanks.
How about the follwoing steps?
- Check gmail inbox every 30 mins
- Classify the pile of email that I got for the last 30 mins into multiple categories (using If-Then I think)
- Summarize selected pieces of email using LLM
- Summarize further as a single report
I am a bit confused how to summarize multiple emails at a time instead of one by one…
It looks fine
What you want to summarize in 3 and 4
And do you need a rough workflow template?
That would be nice if I am given a template. I am trying to figure out how to deploy what I would like to do with n8n.
If I just code in Python, I would say, my python code pulls all the email pieces received for the last 30 minutes, apply a ioop function and store timestamps and contexts of the emails. Then, I apply LLM to each piece of email to summarize what actions needs to be taken if its sales email etc.
Here is a quick template about how to do categorization in n8n for emails.
There are many ways
The first one is based rule. You can write the rules to do it
The second one is based on AI with prompts.
You can see 4 emails is being categoried to promotions or socials.
After that, use AI to summerize the emails then email the report.
1 Like
Thank you so much!
I will try your template.
Please let me know if I am mistaken.
Suppose I set every 30 minutes on Schedule Trigger then this first Gmail node gather all the pieces of email received for the last 30 minutes instead of only one most recent one. Then this AI model takes care of classifiying and summarizing those emails?
Or you can use this. It will trigger every 30 minutes and give you the new emails directly.
1 Like
