EASILY Connect AI Agents to ANY Software | Relevance AI & GPT’s

Easily Connect AI Agents to ANY Software | Relevance AI & GPT’s

Hey everyone! In this quick tutorial, I'll show you how you can connect your AI agents to almost any software quickly and easily, without any complicated API setups or coding. I’ll demonstrate this using the relevance AI platform, but this method also works for OpenAI Assistant or custom GPTs. This is essential because for AI agents to be genuinely useful to most companies, they need seamless integration with their software.

Many find setting up these integrations via APIs a hassle—time-consuming, unclear documentation, coding, and tricky API authorization. The trick I'll show you eliminates all that, making it possible to connect to almost any software within minutes. So if you're serious about building AI agents, stick around, as this will be a huge time-saver.

Starting with Relevance AI

First, I'll show you how to connect agents in Relevance AI:

1. Set up a Lead Manager Agent: This agent will research new leads, generate a company summary and a personalized call script using predefined tools. For example, use a Google Sheets tool to save the research data and a Google Docs tool to store the call script.

2. Connecting to Google Sheets: Instead of using Google's API directly, we use make.com—a no-code automation platform that provides native integration with thousands of software. Here’s how:

   • Create a new scenario with a custom webhook in make.com.
   • Use the webhook URL to post data from the agent.
   • Set up the webhook to receive and then send data (like a company summary) to Google Sheets.

3. Connecting to Google Docs: Similar to Google Sheets:

   • Create an HTML formatted call script.
   • Use the custom webhook to send this script to Google Docs.
   • Create a document with content as the HTML formatted script.

Sending Data Back to Agents

For sending data back to the agent:

1. Set up HTTP Module in make.com:

   • Use Relevance AI's endpoint with a POST request.
   • Include headers for content type (application/json) and authorization (API key).
   • Send specific conversation ID from the agent along with the data.

2. Wait for Agent Availability:

   • Add a short sleep module to ensure the agent is finished processing before it receives the new request.

Triggering Agents from Other Software

Triggering an agent is similar but simpler than sending data back. Set up an HTTP request in make.com without needing a specific conversation ID.

Using OpenAI Assistant and Custom GPTs

Here's how you can use these integrations with OpenAI and custom GPTs:

1. Create Custom Actions in GPTs:
   • Export the tools from Relevance AI and import into OpenAI.
   • Set up API keys and schema in OpenAI's custom actions.
   • Test and ensure the custom actions perform as intended.

You can use similar setups to automate almost any interaction required, such as repurposing YouTube content.

Keywords

Keywords: AI agents, relevance AI, OpenAI, custom GPTs, no-code automation, make.com, API setup, Google Sheets, Google Docs, lead manager, integration.

FAQs

FAQ:

Q1: What is the easiest method to integrate AI agents with other software? A: Using make.com’s no-code platform allows easy integration with thousands of software by simply setting up webhooks.

Q2: How do I send data from a lead manager agent to Google Sheets? A: Use custom webhooks in make.com to post data from the agent. The saved data can then be used to update Google Sheets automatically.

Q3: Can I connect Custom GPTs in a similar way? A: Yes, you can export tools created in Relevance AI to OpenAI and set up custom actions using API keys and schema provided.

Q4: How can I get data back from an AI agent to be notified within Relevance AI? A: Set up HTTP requests with the proper API keys and conversation IDs and use a sleep module to ensure the agent is not busy before sending new data.

Q5: What is the advantage of using make.com over Google’s APIs directly? A: make.com simplifies the process with no complicated API setups, clear native integrations, and easier authorization, making it an efficient choice.