ad
ad
Topview AI logo

I Made an AI Assistant to Handle Emails (Zapier Central)

Education


Introduction

In this article, I will guide you through the creation of an AI system that automates the process of checking your emails and classifying them. This assistant categorizes emails as either generic (those that don't require your reply) or as inquiries from real people. For the latter, it generates draft replies that you can easily review and send.

To create this email assistant, I'll be using Zapier Central. If you haven't tried it yet, I highly recommend giving it a shot. In many ways, Zapier Central represents the next evolution of custom Generative Pre-trained Transformers (GPTs). Its framework shares similarities with what can be found in chat GPT when creating custom GPTs, including a similar interface, logical lines, and the ability for custom instructions and prompt settings.

One major advantage of Zapier Central over traditional GPT Builders is its seamless ability to connect to your online files, such as Google Sheets or Google Docs. This allows you to use these files as a knowledge base. Any real-time updates made to these documents will automatically refresh the bot's knowledge, eliminating the need for re-uploading. This feature is especially valuable for businesses that frequently update spreadsheets and need to pull data for their AI assistants.

Moreover, the ease of connecting tools in Zapier Central sets it apart. You don’t need to deal with complicated schemas, API keys, or code. Instead, you can easily connect various applications by simply searching for the app you want to integrate.

Creating Your AI Email Assistant

Let’s dive into how to create this assistant from scratch, using only natural language, with no coding or API knowledge required!

  1. Getting Started with Zapier Central: Once you're signed up, you'll see an interface similar to chat GPT with two main options: "Create Behavior" and "Add Data Source."

  2. Preparing Your Knowledge Base: First, I prepared a file that will serve as the knowledge base for my mock web agency. Remember, the more detailed your knowledge base, the better your assistant can respond. This file should include details about services, pricing, policies, and additional insights. Since this assistant is intended for handling emails, I've also included response guidelines covering tone, clarity, accuracy, and frequently asked questions. All this content was generated by chat GPT for demonstration purposes.

  3. Adding Data Source: Click on “Add Data Source” and select Google Docs. Choose the document you prepared, and it will be linked to the assistant. Since this is an online document, any updates made will automatically refresh the knowledge base.

  4. Creating Behavior: Now it’s time to create the behavior. Specify custom instructions for the bot, detailing what it should do upon being triggered. Be as specific as possible and break down instructions into multiple steps. Here are the instructions I prepared:

    • Check only emails in the primary inbox; ignore promotions and social.
    • Utilize unread emails, understanding the sender's email address, subject line, and body.
    • Classify emails as either no reply, generic, or real inquiry.
    • Generate a draft for emails classified as real inquiries.
  5. Setting Up Triggers: You can select different triggers for actions, such as a new email in your inbox or a command to the assistant like "check my emails." I chose to schedule the assistant to check for emails daily at 9:00 a.m.

  6. Adding Actions: Next, you need to add actions. With Zapier Central, there’s no complicated schema required. You simply click "Add Action," connect your Gmail, and you’re set. I added actions to find emails and create drafts for inquiries.

  7. Testing the Assistant: Once everything is set, you can test the behavior. After a brief wait, the assistant processes and classifies the emails. It correctly identified which emails needed responses and created drafts for them, saving valuable time for the user. You can see these drafts in your Gmail account.

  8. Finalizing the Assistant: Finally, to ensure this process runs smoothly on a set schedule, go back to the assistant, open the behavior, and toggle it to ‘on.’

Conclusion

This automation can save considerable time with mundane tasks involving emails. The user-friendly design of Zapier Central enhances the experience, making automation accessible for all. I encourage you to try it yourself and see what kinds of automations you can implement!


Keywords

  • AI Assistant
  • Email Classification
  • Zapier Central
  • Knowledge Base
  • Automation
  • Google Docs
  • Draft Replies

FAQ

1. What is Zapier Central?
Zapier Central is a platform that allows users to create automated workflows, integrating various data sources and tools without complex coding.

2. Can I connect my Google Drive or Sheets with Zapier Central?
Yes! Zapier Central allows you to connect online files such as Google Sheets and Google Docs as a knowledge base for your AI assistant.

3. Is prior coding knowledge required to use Zapier Central?
No, Zapier Central is designed to be user-friendly, allowing you to create automations using natural language without coding or API knowledge.

4. How often can the assistant check my emails?
You can set the assistant to check emails at whatever frequency you prefer; for example, I set mine to check daily at 9 a.m.

5. Will the drafts created by the assistant appear in my Gmail?
Yes, any drafts the assistant creates will appear in your Gmail account’s drafts folder for you to review and send.

ad

Share

linkedin icon
twitter icon
facebook icon
email icon
ad