Help Center
How can we help? 👋

Automatically tag outbound email replies using Levity templates

The start to finish guide on building a workflow for email classification with AI

Email management can be a time-consuming task, but with the help of automation software, it can be made easier and more efficient. With the help of Levity’s templates and Flows integrations, your inbox can bet set up to smart-tag and organize your emails in a matter of minutes.

Automatically tag emails based on their content.
Automatically tag emails based on their content.

In this article, we will guide you through the steps of setting up an AI Block from a template, and connecting it to a Levity Flow.

Our pre-designed templates accurately categorize leads as interested or not interested, and can be easily tailored to meet your specific requirements. For even more control, you can also create a custom AI Block with all your own data. Click here for the custom AI Block Guide.

Step 1: Clone out the AI Block Template

Head over to AI Blocks on the left menu, and then click “Show all templates” to open the Template library. Here, we’ll select “Categorize replies to cold outbound emails

Notion image

Congratulations, you just created your own AI solution! Yep, it’s that easy.

Step 2: Make adjustments to the training data and retrain [Optional]

Templates can be added to, edited, and retrained as you need. For example, if you also receive a lot of messages asking for a demo and want to filter those into their own label, you can add additional data to that effect.

To add data, click the “Add text” button and follow the prompts.

If using our spreadsheet import option (seen below), you’ll want to create an Excel or Google sheet with your data - there is a guide on how to structure this file on the popup.

Notion image

If you want to use our Gmail or tool import options, follow the guide here for additional help on that process.

If you add new data or make additional changes, you’ll need to click Retrain for the changes to take effect.

Step 3: See a sample prediction on the Test tab

Head over to the next tab and type in some sample text to see how it will perform.

Notion image

Step 4: Connect your Block to your email using Flows

Click the Connect tab and click the button next to “Connect to a Flow template”. This will open another pop-up wizard, where you can choose your trigger (the action that drives the prediction).

In this scenario we want to make a prediction when we receive a new email in Gmail, so we’ll choose the Gmail to Gmail option.

Notion image

When you select the Flow you want to connect, it will take you through a series of options (seen in the above gif).

Lets break that down…

The next page will ask you to choose your AI Block. Select the Block you wish to connect to the Flow, and click Next:

Notion image

On the next screen you will be asked to authenticate your Gmail account - this is to allow access to connect to your emails. We use a service called as our trusted integration provider for this. Click Next when done.

The next screen allows you to filter the emails classified by Levity, by using Gmail's in-built filter operators. For example, if you only want to send emails from your inbox through your AI Block, you'd type in:inbox in this area. If you're unsure, leave this blank. Click Next.

Notion image

The final step in the process is to match the labels in Levity to the Gmail labels you want to map them to. In this example, we trained our AI Block on similar label names to make it easy to match them:

Notion image

You can click Finish in the bottom right corner of the popup when done with your mapping. Congratulations, you just set up a Flow! Now one final crucial step... turn it on!

Notion image

Step 5: Run a test

Notion image

Send a quick email to yourself (similar to how you used the Test tab) and wait to see the email show up in your inbox and get labeled automatically.

Need some help?

We’re available to provide advice, answer questions, and talk through setup options whenever you’re ready. Click here to read more about getting in touch.

Did this answer your question?