Building AI Agents with Make.com

Introduction

• Overview of Make.com AI agents and comparison to NA10 agents.
• Discussion on pricing: Make.com AI agents incur no additional cost; users cover API usage.

Seven-Step Process to Build AI Agents

Step 1: Determine Need

• Decide if you need an AI agent or a simple automation.
• Questions to consider:
  • Does it involve complex decision-making?
  • Does it coordinate multiple steps intelligently?
  • Does it maintain conversational context?
  • Does it need to work across different platforms?

Step 2: Create the AI Agent

• Start creating an agent in Make.com.
• Use OpenAI's GPT 4.1 for advanced capabilities.
• Configure system prompts with ChatGPT.

Step 3: Choose a Trigger

• Select appropriate trigger for your AI agent (e.g., Telegram, webhooks).

Step 4: Creating Scenarios

• Develop scenarios for tools the AI agent will use.
• Example: Internet search using Perplexity AI, email draft creation via Gmail.

Step 5: Add Tools to the Agent

• Connect scenarios/tools to the AI agent.
• Ensure scenarios are active and set to on-demand.

Step 6: Configure Agent Settings

• Adjust settings such as max output tokens, recursion limits, and history count.

Step 7: Deploy and Test

• Deploy the agent and verify functionality through testing.
• Address any unauthorized errors by checking API connections.

Comparison: Make.com vs. NA10

• Make agents require building a tool library; NA10 integrates tools on the fly.
• Make.com lacks visibility into agent operations; NA10 offers more transparency.
• NA10 agents are currently more mature and easier to build.

Use Cases for AI Agents

• Applicable in customer experience, operations, sales, marketing, finance, and more.

Feedback and Improvements

• Need for better integration of tools in Make.com scenarios.
• Desire for more transparency in agent processes.

Conclusion

• Make.com AI agents are functional and offer powerful automation capabilities.
• Both Make.com and NA10 have their strengths; choice depends on specific needs and preferences.
• Ongoing feedback and improvements are expected for Make.com AI agents.