🔥
12d • General Discussion 💬
Automation Idea:- 2 ways
Same Agent, 2 N8N Ways (Pick Your Style)
Let’s build a simple agent:
“AI Reply Assistant for DMs”
→ Reads any incoming DM
→ Categorizes the message
→ Writes a reply
→ Saves the convo in a sheet
Now here are TWO legit N8N methods to build it — both using nodes, but totally different flows
:-WAY 1 — “Simple Chain” Flow
This is the clean, minimal version. Great for beginners.
Nodes:
1.Trigger Node
Gets new DM (Telegram / Insta / WhatsApp etc.)
2. LLM Node
Asks the model:
what type of message is this?
what should I reply?
(Output = JSON: {category, reply})
3.Set Node
Extracts just the parts you need from LLM output.
4. Google Sheets / Notion Node
Saves: message + category + reply
5. Send Message Node
Sends the reply back automatically.
Why use this?
Fast. Clean. Only 5 nodes. No overthinking.
:- WAY 2 — “Split Brain” Flow
This one is more powerful. You split tasks across multiple LLM nodes for better accuracy.
Nodes:
1.Trigger Node
Same start: incoming DM.
2. LLM Node (Classifier)
Only job:
“What type of message is this?”
Output example: “Sales Inquiry”
3.IF Node
Branch based on category.
4.LLM Node (Reply Generator)
Each branch can have its own tone prompt:
Sales reply
Support reply
Casual chat reply
Lead nurturing reply
Much more control.
5.Google Sheets / Notion Node
Logs data neatly.
6.Send Message Node
Replies with the generated text.
Why use this?
Crazy accurate + customizable + scalable.
:-what flow are you building?
A) Simple Chain (5 nodes, fast)
B) Split Brain (accuracy king)
C) Gonna mix both because I’m chaotic
D) Still choosing my vibe
10
8 comments
Automation Idea:- 2 ways
skool.com/ai-automation-society
A community built to master no-code AI automations. Join to learn, discuss, and build the systems that will shape the future of work.
Leaderboard (30-day)
1
🔥
+2508
2
+1225
3
🔥
+1209
4
🔥
+1193
5
🔥
+787
Powered by