Comment by sunnyb23
24 days ago
Currently working on a project that essentially manages comments on social media channels similar to the current products out there (Sprinklr, Buffer etc) but with a additional layer.
I use AI coupled with Semantic Caching (cost control), NLP and content moderations that would speed up community management for users. It has a manual mode, semi agentic mode and fully agentic. It not only generates responses it contextualises it and understands from your comments who the key contributors are and how you're being perceived.
It can help you identify your promoters (potential customers) and also guide you content that works. It's essentially an evolution from the current community management tools and AI tools which currently focus on content generation. It speeds up community management massively with the advanced filtiration and ML in the backend.
What do you guys think?
So in theory I could hook it up to my LinkedIn account and it would go through people commenting on my posts, run sentiment analysis using a LLM, and identify key supporters or detractors?
Yes exactly if LinkedIn allows me to work with its API like this (its restrictive). This would work with YT, Instragram and X out the box. The key thing is this is AI with human oversight and it learns, gets smarter, the data overtime can turn into sales opp. The agent also adapts based on beacon profiles and optimises when to send comments so it doesnt feel robotic and maximises engagement rate.
I'm a former LI employee so I know some people there in product that may give me a chance but you got it one.
scraping via a browser plugin is not an option? would make you independent from any 3rd party api and the browser already has the users credentials
how would it learn? train/fine tune a custom llm?
1 reply →
[dead]