Build a Customer Support Agent + Ticketing System in 20 min (No-Code)

Key Concepts

• Customer support automation
• Ticketing dashboard
• No-code development
• AI agent
• Community support
• Air Table integration
• Circle community platform
• Perplexity AI
• Webhooks
• Status tracking
• Human-in-the-loop
• Google Gemini 2.0
• API integrations

Customer Support Automation System Overview

The video demonstrates a no-code system for automating customer support within a community, specifically built for a Circle community but adaptable to platforms like School, Discord, Slack, WhatsApp, Facebook, email, or websites. The system uses an AI agent to answer tech questions by searching past community posts and leveraging Perplexity AI. The entire process is managed through an Air Table ticketing dashboard that automatically updates statuses.

System Demo and Functionality

1. Trigger: A user posts a tech question in the Circle community.
2. Ticket Creation: The system captures the post via webhook, creates a new ticket in Air Table, including the post title, details, user email, username, status (set to "new"), topics, post URL, and user ID.
3. AI Agent Response: The AI agent searches for similar past questions and answers within the community. If no relevant information is found, it uses Perplexity AI to generate a response. The agent then posts a comment back to the original post with the answer, confidence level, and links to relevant resources.
   • Example: A user asks about connecting Ampify to Relevance AI. The AI agent provides steps, including getting Ampify API details, using Relevance AI's API call, and configuring the API call. It also links to past posts on the same topic.
4. Air Table Dashboard: The Air Table dashboard tracks all tech questions with statuses like "Not Replied," "Replied," "Solved," and "Urgent Resolution."
5. Status Updates:
   • Users can manually mark a post as "Solved."
   • An automated process checks the status of the post and comments to determine if the issue is resolved. It analyzes user comments to see if the question has been answered.
   • The system updates the Air Table ticket status automatically based on the analysis.
6. Post Update: If the post is marked as "Solved" (either manually or automatically), the system updates the Circle post title to include "Solved."

Step-by-Step Breakdown of the Automation Flows

1. Question Submission and Initial Response

1. Webhook Trigger: A webhook in Circle is triggered when a new post is added to the tech help section. The webhook is configured to allow both POST and GET methods.
2. Get Circle Post: The system uses the post ID from the webhook to retrieve the post details from the Circle API.
3. Create Air Table Ticket: The system creates a new record in Air Table with the post details.
4. AI Agent Interaction:
   • The system passes the post title, body, and relevant topics to the AI agent.
   • Prompt: The AI agent is instructed to act as an official assistant, answer questions accurately, and provide responses in HTML format.
   • Workflow:
     1. Use the search tool (Circle API) to find similar past posts.
     2. If search results are insufficient, use the Perplexity AI tool.
     3. Provide a confidence level for the answer (High, Medium, Low).
   • Tools:
     • Google Gemini 2.0 (for general knowledge and reasoning).
     • Circle API (for searching past community posts).
     • Perplexity AI (for additional information and resource linking).
   • The AI agent posts the response back to the Circle post using the Circle API.

2. Status Check and Resolution

1. Air Table Automation Trigger: When the "Check Status" checkbox is marked in Air Table, an automation is triggered.
2. Run Script: A script is executed to trigger a webhook with a specific action ("check status").
3. Get Circle Comments: The system retrieves all comments from the Circle post using the Circle API.
4. Data Processing:
   • The comments are split out and cleaned using HTML to Markdown conversion.
   • The username and comment text are mapped for each comment.
   • All comments are aggregated into a single text string.
5. AI Analysis:
   • The post title, description, time since posted, and aggregated comments are passed to an AI agent.
   • Prompt: The AI agent is instructed to analyze the comments and determine if the issue has been resolved. It should provide a recommendation for the next resolution step and identify the current status.
   • Status Options:
     • Solved: The issue is clearly resolved.
     • Not Replied: No comments or only the initial AI agent comment.
     • Urgent Resolution: The issue remains unresolved for one day or more without replies.
     • Replied: Comments exist beyond the initial AI agent comment.
6. Air Table Update: The AI agent's determined status is updated in the Air Table record.

3. Post Marked as Solved

1. Air Table Automation Trigger: When the status in Air Table is changed to "Solved," an automation is triggered.
2. Run Script: A script is executed to trigger a webhook with the action "solve ticket."
3. Update Circle Post: The system posts a comment to the Circle post indicating that it has been marked as resolved and updates the post title to include "Solved."

Key Arguments and Perspectives

• Efficiency: The system aims to provide faster answers to tech questions within the community.
• Scalability: The no-code approach allows for easy adaptation and scaling of the support system.
• Automation: Automating status updates and initial responses reduces the manual workload for community managers.
• AI-Powered Support: Leveraging AI agents and knowledge bases improves the quality and consistency of support.
• Customization: The system can be customized to fit different community platforms and specific use cases.

Notable Quotes

• "This agent searches through past community questions and uses perplexity to resolve issues."
• "The template will be completely for free in the link in the description below."
• "We want to help people inside of my community as fast as possible with their tech questions."

Technical Terms and Concepts

• Webhook: An automated HTTP request triggered by an event in one system (e.g., a new post in Circle) to another system (e.g., N8N).
• API (Application Programming Interface): A set of rules and specifications that software programs can follow to communicate with each other.
• No-code: A development approach that allows users to create applications without writing code.
• AI Agent: A software program that uses artificial intelligence to perform tasks autonomously.
• Ticketing System: A system for tracking and managing customer support requests.
• HTML (HyperText Markup Language): The standard markup language for creating web pages.
• JSON (JavaScript Object Notation): A lightweight data-interchange format.
• Markdown: A lightweight markup language with plain text formatting syntax.

Logical Connections

The video presents a system where each component is logically connected to the others. The webhook triggers the process, the AI agent provides the initial response, the Air Table dashboard manages the tickets, and the status updates ensure that issues are resolved efficiently. The use of AI for both answering questions and determining resolution status creates a closed-loop system that minimizes manual intervention.

Data, Research Findings, or Statistics

The video doesn't present specific data, research findings, or statistics. It focuses on demonstrating the functionality and architecture of the automated support system.

Synthesis/Conclusion

The video provides a detailed walkthrough of a no-code customer support automation system built for a Circle community. The system leverages webhooks, AI agents, and Air Table integration to streamline the support process, provide faster answers, and reduce manual workload. The architecture is adaptable to other community platforms and use cases, making it a valuable resource for anyone looking to automate their customer support or community management efforts. The key takeaways are the importance of AI-powered support, the efficiency of no-code development, and the value of a well-integrated ticketing system.