John A

AI-Powered LinkedIn Content Automation

Nov 2024 - Present • 1 yr 10 mos

I built an automated content publishing workflow that connects Google Sheets, n8n, ChatGPT, and the LinkedIn API into one end-to-end system. Google Sheets was used as the content management layer, where content ideas and information were stored and organized. I configured n8n to monitor the sheet and identify entries ready for processing. When new content was detected, the workflow retrieved the relevant information and passed it to an AI Agent powered by ChatGPT. I configured the AI stage to process the raw information and generate structured, LinkedIn-ready content based on predefined instructions. The generated content was then returned to n8n, formatted, and prepared for publishing. For the publishing stage, I integrated the LinkedIn API so the workflow could send the generated content directly to the required LinkedIn destination. After processing, the workflow returned to Google Sheets and updated the corresponding row with the publishing status, creating a simple tracking system and reducing the possibility of duplicate processing. This project gave me practical experience with AI content generation, API integration, data processing, workflow orchestration, automated publishing, and status tracking using n8n. Tools: n8n, ChatGPT/OpenAI, Google Sheets, LinkedIn API

AI Voice Agent with Conversation Memory

Nov 2024 - Present • 1 yr 10 mos

I built an AI voice agent workflow using Make.com, an AI voice platform, and Airtable to handle incoming calls while maintaining caller information and conversation context across multiple interactions. The workflow begins when a caller contacts the voice agent. The agent answers the call and processes the conversation in real time, allowing callers to ask questions and interact naturally. During the interaction, relevant caller information can be captured and passed through the automation for storage and future use. I configured Airtable as the memory layer for the system. When a new call comes in, the workflow checks the caller's phone number against existing Airtable records. If the caller has interacted with the agent previously, the system retrieves the stored information so the agent can recognize the caller and use the available context during the conversation. For new callers, the workflow can create a new record and store relevant information for future interactions. This allows the system to maintain a history instead of treating every incoming call as a completely new conversation. Building this project strengthened my experience with AI voice agents, persistent memory, database lookups, caller identification, automation logic, and real-time communication workflows. Tools: Make.com, AI Voice Agent, Airtable, APIs