VIDEO
HIGHLIGHT
Log InSign up
Claude Code Just Made Building Voice AI Agents Stupidly Simple!
SummaryTranscriptChat

Claude Code Just Made Building Voice AI Agents Stupidly Simple!

Building a Voice AI System with Claude Code

Overview of the Voice AI System

  • The speaker describes building an entire voice AI system using Claude code without manual setup, simply by prompting Claude.
  • The system is designed for Retail AI and includes functionalities like checking calendar availability, booking appointments, and post-call routing.
  • The speaker emphasizes that no coding background is required to build these systems; all necessary prompts will be provided for free.

Features of the HVAC Voice Agent

  • The voice agent created is specifically for John's HVAC company, capable of handling service appointments and extracting caller information automatically.
  • It builds three unique workflows: one for checking availability, another for booking appointments, and a third for post-call webhooks.

Getting Started with Cloud Code

  • Instructions are given on installing Cloud Code from cloud.ai and ensuring users are on the pro plan ($20/month).
  • Two methods to prompt Claude are introduced: a comprehensive prompt or a step-by-step approach. Links to resources are provided in the description.

Accessing Resources

  • A free community classroom section offers additional resources including prompts specifically tailored for HVAC voice systems.
  • There’s also mention of a paid certification program in partnership with Retail AI to become certified in building client-ready voice agents.

Building the Voice Scheduling System

  • The speaker decides to use a comprehensive prompt method to create the HVAC scheduling system.
  • Users need to create a folder on their local computer before pasting the prompt into Cloud Code. This ensures proper organization of files related to the project.

Creating an HVAC Workflow with Naden API

Initial Setup and Greeting the Caller

  • The agent should greet the caller and inquire about their issue, specifically regarding AC, heating, maintenance, or other concerns.
  • Essential information to collect includes the caller's name, phone number, service address, and whether the issue is urgent.

Booking Consultation and Function Calls

  • Offer to book a consultation while utilizing real function calls via Naden instead of mock functions.
  • Emphasizes the importance of creating workflows through the Naden API rather than using JSON files to avoid manual copying and pasting.

Integrating Google Calendar and Notifications

  • Use Google Calendar for checking availability and creating events when booking appointments; also expose webhook endpoints for various functions.
  • If an issue is marked as urgent, a Slack message should be sent; if an appointment is confirmed, an email notification will be dispatched.

API Key Configuration

  • The process requires entering several API keys: retail AI API key, Naden cloud URL, and Naden API key.
  • Instructions are provided on how to create these keys within the Retail AI platform for seamless integration.

Finalizing Agent Creation

  • Users are guided through setting up their Retail AI account if they do not already have one; this includes navigating to settings to generate necessary API keys.
  • Once all configurations are complete, it confirms that the Naden API is connected successfully and begins creating the HVAC workflow with required webhooks.

Workflow Creation and Integration in Retail AI

Initial Setup and Permissions

  • The process begins with the user allowing necessary permissions for the system to function properly, which is crucial as multiple permission requests will occur.
  • A workflow named "naden" is created, activating web hook URLs essential for further integrations.

Workflow Components

  • Three distinct workflows are established: one for checking availability, another for booking appointments, and a third for post-call web hooks.
  • The system automatically manages production URLs without requiring manual changes from test to production settings.

Google Calendar Integration

  • Users must manually select their Google Calendar after creating credentials; this step cannot be automated by the system.
  • The second node focuses on booking appointments, emphasizing the need to attach Google credentials correctly.

Post Call Web Hook Functionality

  • The post-call web hook analyzes data to determine urgency; urgent matters trigger Slack notifications if configured.
  • Confirmation emails are sent upon appointment confirmation, but users may need to adjust formatting or content as needed.

Community Resources and Learning Opportunities

  • For community members, a comprehensive voice AI prompt guide is available that aids in creating prompts for cloud code applications.
  • Emphasis on understanding systems thoroughly is highlighted; while automation simplifies setup (80% done), complete responsibility lies with the user when preparing production-ready solutions.

Creating a Customer Service System with Cloud Code

Overview of the System Prompt

  • The system prompt defines the role of a friendly and professional customer service representative for Cool Flow, focusing on greeting customers and checking availability using specific functions.
  • Integration with Naden is highlighted, showcasing how information is sent to a webhook seamlessly, demonstrating the efficiency of this setup.

Functionality and Data Handling

  • The process allows users to modify prompts by communicating directly with Claude, enabling adjustments if certain steps are missing or need refinement.
  • While Claude generates about 80% of the necessary code, user interaction is required to finalize details and troubleshoot issues effectively.

Advanced Features and Data Extraction

  • Users can switch models (e.g., GPT 5.2 or GPT real-time), noting that more powerful options may incur higher costs but offer enhanced capabilities.
  • Post-call data extraction retrieves essential information like call summaries and caller details, which can be integrated into CRM systems or communicated via Slack for urgent matters.

Community Engagement and Future Developments

  • The speaker encourages community feedback on Cloud Code features and expresses intent to create exclusive content focused on building voice agent systems.
  • Emphasizing the ease of creating agentic workflows without manual node connections, the discussion points towards an innovative future in voice AI systems.
Video description

🚀 Full Voice AI Course + Certification 👉 https://www.skool.com/aiworkshop Prompts for this video + free resources 👉 https://bit.ly/aiw-lite-about Sign up for a free n8n cloud account! 👉 https://n8n.partnerlinks.io/39tj2ld0gxo6 Sign up for a free Retell AI account! 👉 https://dashboard.retellai.com/?ref=aiworkshop Work with us to Implement Voice AI Into Your Business 👉 https://bit.ly/AIW-Discovery-Call In this tutorial, I show you how to build a complete voice AI system using Claude Code, Retell AI, and n8n, all by prompting Claude instead of manually wiring workflows. You’ll see how Claude Code can create a full voice AI agent in Retell AI, generate real functions, handle post-call analysis, and automatically build multiple n8n workflows for checking calendar availability, booking appointments, and routing calls based on urgency. This walkthrough covers how to connect voice AI with Google Calendar, Slack, and email, and how to assemble a real, client-ready AI system step by step — even if you don’t have a coding background. If you’re interested in voice AI, AI agents, n8n automations, or building sellable AI systems for businesses, this video shows how the process is changing with Claude Code. Video Chapters: 0:00 Intro 0:52 Designing the Voice AI Agent (What We’re Building) 2:10 Writing the Claude Code Prompt (Full System Build) 8:08 Connecting Retell AI + n8n via API to Claude Code 13:14 n8n Backend Automations (Scheduling, Routing, Logic) 18:01 Retell AI Agent Walkthrough (Prompts, Functions, Post-Call Data)