A social media automation platform needed to act on comments at scale. Brands receive hundreds of comments per post, and real buying signals get buried under praise, spam and off-topic replies. Sorting them by hand was slow and good leads were being missed.
What I built
An LLM-powered module (Claude) that reads incoming comments and classifies their intent, for example buying interest, questions, complaints or spam
Integration with the platform's social data feed (SocialAPI.ai) and CRM (GoHighLevel), so each classified comment can trigger the right follow-up automatically
A versioned technical specification that keeps the classification logic documented and auditable as it evolves
Tech stack
Claude, SocialAPI.ai, GoHighLevel, webhooks and REST APIs.
My role
Designed the intent categories and classification logic, wrote the technical spec, and built the integration together with a collaborator.
Client details are kept confidential under NDA.
Like this project
Posted Sep 28, 2026
An LLM module that sorts social comments by intent and routes buying signals into the CRM, so leads stop getting lost in the noise.