Reads comments, finds the right event details, and replies in the brand's voice, with a safety filter for spam and aggressive comments.
- Instagram comment
- Understand intent
- Find event details
- Safety check
- Reply or escalate
The problem
What needed to be handled
The team could not keep up with Instagram comments asking about events, and risky comments had to be caught before any reply went out.
What we built
System details
- An n8n workflow that reads Instagram comment events and works out what the person is asking.
- Looks up matching events with semantic search, then writes a human-sounding reply.
- A safety layer (Lakera) that detects spam and aggression before replying.
- Escalation logic for support cases.
Facts
- Safety layer
- Spam and aggression detection built in
- Replies
- Automatic
Stack
- n8n
- Instagram Graph API
- OpenAI
- Supabase Vector Store
- Lakera AI
Related services
Ready to put a system in place?
Let's discuss your calls, leads, and workflows. Get in touch to plan your next step.