• NanoBits
  • Posts
  • Gmail + Make.com + ChatGPT: Summarise unread newsletters in 7 steps

Gmail + Make.com + ChatGPT: Summarise unread newsletters in 7 steps

Nanobits Product Spotlight

EDITOR’S NOTE

Dear Nanobiters,

Finally, it’s weekend!
It's time to sit down, relax, and catch up on all those unread newsletters you have been meaning to read over the week.

You start reading one and realize midway that you are not interested in the content. Then you start another, only to realize that this newsletter was also the least interesting.

Then, you finally give up and start binging that new episode of Squid Games on Netflix!

Happens to me every 3 out of 4 weekends!

The flood of newsletters was interfering with my ability to read them effectively. Searching for worthwhile content left me fatigued, often wasting precious time just sorting through the mess.

They gathered dust in my inbox like old newspapers, their unread indicators silently demanding attention.

I also tried a few email summarizer tools, like Meco. They were expensive and did a shoddy job of summarizing the newsletters.

So, I built a newsletter summarizer using Make.com and my favorite, ChatGPT.

The best part: It’s free and needs no coding!

In this newsletter, I will guide you to build the same summariser in 7 easy steps.

It took me less than 30 minutes to set up the whole workflow.

So, let’s get started.

ARE YOU NEW TO MAKE.COM?

It’s a super easy and intuitive tool to learn.

Here’s a beginner-friendly guide/primer I wrote a few weeks ago. It will help you get started with Make.com.

How to create a Make.com workflow that summarizes your unread newsletters

This simple automation will summarize your unread newsletters and email you the summary. You don’t have to scroll through all your emails to find that one article you want to read.

Let’s begin.

Click on scenarios on the left-hand side.

Click “Create a new scenario” on the top right corner of the page.

This now drops us into the scenario editor or designer, where we can build our automation.

Module 1: Google Docs

In the center of the screen, you'll see this large plus icon. This is the first step of our scenario. We will start by adding a Google Docs module.

A few things to note:

  • If you haven’t used Make or any Google modules before, you must sign in to your Google account to set up a connection.

  • The “now“ timestamp in the newsletter name determines when the newsletter was created.

  • Make requires you to put some content in to create a Google Doc. Since we want a blank document, I have added one hyphen so it’s not empty.

Module 2: Email

This is the most tricky part of the workflow, so pay close attention to this section.

To connect restricted Google services, like Gmail and Google Drive, to Make, you must create a project on the Google Cloud Platform and a custom OAuth client.

You can follow this guide with additional required steps to connect.

After connecting your Gmail account, choose which newsletter folder you want summarized.

For this newsletter, I have chosen 5 unread newsletters for one operation. Depending on how frequently you want these summaries sent to your inbox, you can choose less or more.

I have choosen The Rundown AI for this summary because it collates the most significant news in the AI space and delivers them to your inbox everyday. 

I usually don't have the time to catch up with the AI news everyday, so I read all 6 of them on the weekends. 

Now, that's just too much information!

Module 3: Newsletter Summary Writer [ChatGPT]

Now, let’s add a ChatGPT module to our flow. This module will send the unread newsletters from the chosen email folder to ChatGPT for processing.

To connect Make and ChatGPT, you need an OpenAI API key, which you can find here.

Then click on the plus sign on the right of your Email Module to add the ChatGPT module.

Note: Choose Completion (Prompt) (GPT and o1 Models) trigger from the list of actions. 

Here is the prompt I used for the module. You may modify it according to your preference.

As an expert newsletter writer, your task is to analyze the content provided in the user prompt and select the 5 most significant and engaging stories. You will then reword these stories into bite-sized segments, without adding any additional information. After each reworded story, remember to leave 2 line breaks. If there are links available in the original story, make sure to include them. This task requires a keen eye for detail and the ability to summarize stories effectively while maintaining their essence and relevance.

💡 Please note that if you set the maximum number of tokens to use in the completion (output). If empty, the limit of the model will be used.

I prefer to leave it empty.

Module 4: Collates News Summary in a Google Doc

In this module, we will collate all the news summaries that ChatGPT created in the Google Docs that we created in the first module based on the unread newsletters in the email folder.

Array Aggregator

Since I have more than 1 email to process, I used an array aggregator. This tool waits for all the information from the unread newsletters to be added to the Google Doc before proceeding to the next steps.

Array Aggregators convert multiple bundles of data into a single array containing collections of specified items.

Please note that the source of the array aggregator is the email module. If you are confused ask yourself: At which point in this scenario does your aggragting process start?

The array aggregator acts as a one-way operation - you cannot directly access the data inside it. To retrieve the aggregated summary data, you'll need to add a second Google Doc module that pulls the results from the first Google Doc module where the aggregation result was stored.

Module 5: Collates News Summary in a Google Doc

Module 6: Expert Content Curator [ChatGPT]

You don’t have to recreate a connection between Make and ChatGPT this time. Select the Completion (Prompt) (GPT and o1 Models) trigger from the list of actions.

Here is the prompt I used for the module. You may modify it according to your preference.

You are an expert content curator with over a decade of experience in analyzing and summarizing newsletters, particularly those focused on technology, AI, and major industry players. Your specialty lies in identifying the most impactful stories, simplifying complex information, and presenting it in an engaging, digestible format. You have a knack for capturing the essence of each story while maintaining a lively and approachable tone.  
Your task is to analyze the newsletter content provided in the user prompt and select the 7 most significant and noteworthy stories. Focus primarily on AI tools, large companies like OpenAI, Microsoft, Apple, X, Elon Musk, and groundbreaking AI breakthroughs. Avoid overly technical details or investment-related content. Reword these stories into concise, bitesize segments, ensuring each one is energetic and easy to understand. Include one emoji per story to add a touch of personality and leave two line breaks after each segment. If available, include the associated link below the reworded story.  
Here’s what you need to keep in mind:  

Prioritize stories that are relevant, impactful, and align with the focus areas mentioned above.  
Reword the content in a way that is clear, engaging, and free of unnecessary jargon.  
Do not add any information that isn’t explicitly present in the original content.  
Maintain a consistent tone throughout, avoiding robotic or overly formal language.

For example, if the original story is:
"OpenAI has announced a new AI model that can generate high-quality video content in seconds, revolutionizing the creative industry."

You might reword it as:
"OpenAI just dropped a game-changing AI model that creates stunning videos in seconds! 🎥 This could totally transform how we think about creativity. [Link]"  

Now, analyze the following newsletter content and provide the 7 most significant stories in the requested format.   

Module 7: Publish the summary to your INBOX

You can share the summaries by emailing them to your inbox, posting them on a Slack channel, or saving them in another document.

For this newsletter, I have chosen to email them to my inbox.

Now, let’s examine our progress so far.

These are the 5 unread newsletters in my folder.

Here’s the result of one iteration of the workflow.

The workflow is now ready!

The next step is to schedule it.

I don't want to come here and manually run this every weekend.

To turn scheduling on, press this toggle in the bottom left-hand corner. Then, you can configure the schedule to run on certain days of the week or month.

Remember that the more often it runs, the more operations your scenario will consume. So, think of a reasonable interval at which to run this.

And just like that, we've saved some precious time by automating repetitive tasks.

End Note

Reading newsletters should not consume your entire day. You can create folders for specific newsletters of your choice and make as many summarizers as you want.

No more drowning in unread newsletters - now you can quickly scan the key points and decide what's worth your full attention.

The setup might take a few minutes, but the time you'll save in the long run is invaluable.

I'd love to hear about your experience with this workflow. If you found it helpful, consider sharing it with others drowning in newsletter overload, and subscribe to Nanobits for more productivity automation tips.

Do you have questions about setting up your own newsletter summarizer?
Reply to this email. I'm here to help troubleshoot and offer guidance.

Until next time, happy reading – the smarter way!

P.S. Remember to periodically check your automation to ensure it's running smoothly and adjust the prompts as needed for better summaries.

Share the love ❤️ Tell your friends!

If you liked our newsletter, share this link with your friends and request them to subscribe too.

Check out our website to get the latest updates on AI.

Reply

or to participate.