Wish you had a personal team of researchers? This tutorial will show you how to build a powerful automation that turns any article or webpage into ready-to-post social media content. Just email a link, and watch as AI does the heavy lifting - researching the content and crafting platform-perfect posts for your social channels.

Here's how it works:

1. You email a link to your automation
2. Perplexity AI analyzes the content and extracts key insights
3. ChatGPT transforms those insights into platform-optimized posts
4. The content automatically posts to your chosen social media channels

While we'll use LinkedIn as our primary example, the beauty of this system is its flexibility – you can add X (Twitter), Facebook, or any other social platform you prefer. The process remains the same for each channel, so feel free to expand the automation to as many platforms as you need.

Setting up your Perplexity account

We are going to be using Perplexity AI to conduct research on our topic/article and provide a clear summary of the results. Perplexity AI is like having a super-smart research assistant in your pocket. Think of it as a modern search engine that doesn't just find information – it actually understands and explains things to you in a conversational way.

1. Go to Perplexity and create an account - we will need a paid account to get access to the API. 
2. Go to API settings and generate an API key. Name the API key “Make.com”.
3. Add some credits to your account - $5 dollars will be plenty
   ‍

4. Make a secure record of your API key as it will be required to connect our AI to the Make scenario later. Consider storing it in a password manager. 

Tip: Treat this key like you would treat a password. It grants access to your account and must be kept secure. 

Setting up your OpenAI account

In order to use AI to generate our post content we need to set up a paid OpenAI account. To do this go to OpenAI and create an account, ensuring you add a payment method since using the API is a paid feature. 

1. Go to OpenAI and create an account
2. Go to the API Key page and generate an API key. Name the API key “Make.com”.
3. Make a secure record of your API key as it will be required to connect our AI to the Make scenario later. Consider storing it in a password manager. 

Tip: Treat this key like you would treat a password. It grants access to your account and must be kept secure. 

‍

‍

Setting up the AI assistant in OpenAI

Once you write your content, we are going to send it to OpenAI for repurposing. In order to do this we need to set up several AI assistants in OpenAI. The AI assistant is just like using ChatGPT through the web but with the added feature of pre-programming GPT to respond exactly how we want it to. 

We are going to be creating a bot for LinkedIn. However, you can repeat the process to create a bot for any social media platform you’d like to automate! 

Step 1 - Writing our prompt

Let's use ChatGPT to create our AI assistant's prompt - after all, who better to write AI instructions than AI itself?

1. Head over to ChatGPT.com and start a new conversation. 
2. Ask ChatGPT to do the following: 

Write a prompt for an LLM assistant. When I provide it research from Perplexity AI, it will restyle the content for LinkedIn

Tip: You can swap out “LinkedIn” for any social media platform

‍

‍

3. Copy the output to your clipboard, we will need it soon. 

Step 2 - Train Content Bot

1. Navigate to the Assistants menu and create a new assistant. 
   ‍

2. Name the Assistant “LinkedIN Bot”.
3. Add the prompt you just copied to the System Instructions field. 
4. Repeat this process for the different social media platforms you’d like to automate. 

That’s it for OpenAI, our assistants are now ready to be called via the API and integrated into the Make scenario. 

Configuring the Make scenario

With our Perplexity and OpenAI accounts configured, we can start automating the process. To do this, we will be using Make.com, a powerful iPaaS middleware which allows you to automate almost anything.

Step 1 - Create a new Scenario

Go to Make.com and sign up for a free account if you don’t have one already. The free account will be sufficient while you test out the automation. You may wish to consider upgrading once you have everything working correctly. 

‍

‍

Step 2 - Name Scenario and add Trigger

1. In the top right corner of the scenario, change the name to “AI Researcher"
2. Click the Plus symbol and search for the Webhooks module, select “Custom mailhook” as the trigger. 
   ‍

‍

3. Add a new Mailhook and name it AI Researcher 
   ‍

‍

Step 3 - Add Perpexity AI researcher

1. Add a new module, search for “Perplexity AI” and select “Create a chat completion” as the action 
2. Add a new connection using the API key you saved from Perplexity. 
   ‍

3. Select the latest model from Perplexity AI - play around with the different options to see which will get you the best results at your price point. For this tutorial we’ve gone with Mistral-7B-instruct 
   ‍

‍

4. Add the following Prompt into the content field - we have included the make.com mapping directly in the prompt to make things easier for you. 

‍

Analyze this {{1.text}} and provide:

Key Stats:

Most surprising numbers/data

Standout metrics

Growth trends

Main Insights (Top 3):

Key takeaways

Unique perspectives

Future implications

Content Potential:

Best social media angles

Most shareable elements

Attention-grabbing hooks

Credibility Check:

Source reliability

Data verification

Timeliness

Focus on finding: ✦ Counter-intuitive findings ✦ Practical applications ✦ Industry impacts ✦ Supporting evidence

Please highlight anything that seems outdated, biased, or unsupported by data.

5. Select “User” as the Role. 

Step 4 - Add a Router

1. Add a new “Router” module and create three branches by clicking on the module once after creation.
   ‍

‍

Step 5 - Add our AI Assistant

1. On the 1st branch, Add an OpenAI module with the action “Message an Assistant”.

2. Create a connection. Paste the OpenAI API key you saved earlier into the API field. You can leave Organization ID blank.

‍

‍

3. Select the LinkedIn Bot from the Assistant Dropdown

4. Map the perplexity output into the message field. You might need to expand Choices and Messages Arrays to access the correct field. 

‍

‍

5. Repeat this for the other GPT AI Assistants that you have created on the other two branches
   ‍

‍

Step 5 - Connect our Researcher to LinkedIN

1. On the 1st Branch, add a new module. 
2. Select the LinkedIN module and “Create a User Text Post” as the action
3. Connect your LinkedIn Account by signing in. 
   ‍

4. In the Content field, Map result from ChatGPT

‍

5. Repeat the process to connect your preferred Social Media Channels on the other branches. 

Step 6 - Switch your scenario ON 

1. At the bottom of the Scenario builder, switch your scenario ON
2. Go to the first Module and copy the Mailhook address - this is where we will send our articles
   ‍

‍

Using your AI researcher

To let our AI researcher know that we have a new topic for it to unpack, we need to send it an email with the link we want it to unpack and post to social media. 

1. Go to your email client and start a new message. 
2. Paste the address we just copied from Make - you might want to save this as a contact in your Email client
3. Write “Research” in the Subject
4. Paste the link to the article/post of story you want your researcher to post about (one article per email)
   ‍

5. Click send! The post will appear on LinkedIn in a few minutes. 

‍