🚀 Master AI Automation: 10 Levels from Basics to Advanced

Introduction

In today's busy world, knowing AI automation isn't just nice to have – it's a must for getting more done efficiently. If you're a beginner want to enhance your skills, this guide is here to walk you through every step. We'll start with the basics of prompt engineering and work up to more complex AI workflows.

Level 1: Prompt Engineering Framework

1. What is Prompt Engineering?

Prompt engineering is like asking the right questions to get the best answers from AI Chatbot like ChatGPT or Claude. It's about creating clear and simple instructions, so the AI understands exactly what you need. Think of it like giving directions to a friend: the clearer you are, the better they can help you.

2. Basic Framework for Creating Prompts

To make effective prompts, follow these three steps: Purpose, Clarity, and Improvement.

1. Purpose

   Purpose is about knowing why you are making a prompt. It’s like having a goal or reason for doing something.

   

2. Clarity

   Clarity is about making your prompt easy to understand. You can add a line like, "Ask me for more details if needed". This way, if your AI is unsure, it can ask for more information instead of guessing. This ensures better accuracy and clarity in the AI's responses, improving overall communication and results.

   

3. Improvement

   Improvement is about making your prompts better over time. You can ask the AI “How can this be better?” to make its response better. Your AI will then review its own work and suggest improvements, like a perfectionist friend who loves giving advice.

   

3. Example of a Simple Prompt

Let's walk through the final prompt using this framework.

Final Prompt

List some ways we can use virtual assistants in my AI automation agency. I want to increase profits, cut costs, and spend less time on low-value tasks. Ask me for more details if needed.

Here’s what you might get:

Level 2: Automating Social Media Posts

1. Introduction to Social Media Automation

Managing multiple social media accounts can be a lot to handle. Automating your posts simplifies the process, saving you time and effort. This lets you focus on creating engaging content and interacting with your followers. By using automation tools, you can keep a consistent online presence, schedule posts in advance, and stay on top of timely updates. This way, you can devote more energy to building your brand and connecting with your audience.

2. Tools

ChatGPT and Claude: These AI tools can quickly generate engaging posts based on the information you provide. They're like smart, creative partners who help you keep your social media fresh and captivating, making your content stand out effortlessly. With their help, you can focus more on connecting with your audience while they handle the content creation.

3. Setting up Your Prompt

• Here’s a basic prompt structure you can use:

Act as a Twitter expert. Your task is to convert the following article into a concise, engaging tweet. The tweet should: 

- Be clear and focused on one key message.

- Use a hook to grab attention. - Include a brief explanation of the article's key point. 

- End with a thought-provoking question and relevant hashtags. 

- Stay within 240 characters. 

Article: [Paste your article text here]

4. Example Workflow:

• Step 1: Find relevant content or news articles. For example, search for the latest updates on AI.

• Step 2: Use ChatGPT to craft a tweet based on the article. Give the AI the article text and specify your requirements.

• Step 3: Refine the tweet by asking the AI for improvements.

• Step 4: Use a social media management tool to automate posting the tweet.

Level 3: Coding within No-Code Software

1. Introduction to No-Code Tools

• Make is a no-code platform that helps you automate tasks and connect different tools and services easily. It supports adding custom code elements for more complex automations.

• You can use Python code in Make to enhance its capabilities, going beyond what typical no-code tools offer.

2. Setting Up a Simple Automation in Make

Let’s automate a simple task, like getting text from a YouTube video and turning it into a transcript.

Step 1: Sign Up and Setup

• Sign up and create a new scenario.

  

Step 2: Add Modules

• Click on the plus icon and add modules. For this task, use 6 modules to get the transcript of a YouTube video.

  • First, add the “Tools” module and choose “Set variable.”

    

    • Fill in “url” in the “Variable name” box.

    • In the “Variable value” box, paste the link to your favorite YouTube video.

      

  • Add the “HTTP” module and choose “Make a request” to get the HTML from the video page.

    

    • Click on the HTTP icon, and the following box will appear.

    • Just click on the “URL” box and choose “url”.

    • Choose "GET" for the method. (The other fields will be automatically filled, so don’t worry about them)

      

  • Add the “Text Parser” Module and choose “Match pattern” to find the transcript link.

    

    • In the 'Pattern' box, enter the following exactly as shown:

      (https\:\/\/www\.youtube\.com\/api\/timedtext[\S][^"]*)

      

    • In the “Text” box, choose “Data”, which is the result of module 2.

      

  • Add another “HTTP” module, and choose “Make a request” to get the transcript.

    

    • In the “URL” box you will fill in:

      ($1 is a variable resulting from module 3)

    • Choose “GET” for the method.

    • These actions help to fetch the actual transcript from the extracted URL after replacing encoded characters.

      

  • Add the “Text Parser” module and choose “HTML to text”.

    

    • In the “HTML” box, select “Data”, which is the result of module 4.

      

  • Add the “Google docs” module, and choose “Insert a Paragraph to a Document”.

    

    • First, you need to connect with your Google account by clicking on “Add”.

      

    • In the “Document ID box”, click on your Google Docs file.

      

    • In the “Appended Text” box, choose “Text”.

      

Step 3: Run the scenario

• When you successfully set up the 6 modules, click on “Run once”. The transcript of your favorite video will automatically appear in your chosen Google Docs file.

  
  

Level 4: Using RSS Feeds for Content Automation

Now, we focus on automating the collection of content from multiple sources using RSS feeds. RSS (Really Simple Syndication) feeds keep you informed of new updates from websites and blogs. Here's a detailed guide on setting up and using RSS feeds to streamline content collection.

1. What is RSS Feed?

RSS (Really Simple Syndication) feeds enable you to automatically receive updates from websites and online sources. Whenever new content like blog posts, news articles, or tweets is published, it's delivered directly to you. This tool saves you time by eliminating the need to manually check each website for updates, ensuring you always have the latest information at your fingertips.

2. Setting Up Your RSS Feed

• Step 1: Create an RSS Feed Account:

  • Go to an RSS feed aggregator service like RSS. Sign up for an account if you don't have one.

• Step 2: Add RSS Feed:

  • Search for the topic you want to explore. For example, you pick “Technology” field, then click on "Add Feed". Initially, you get to add up to 15 feeds for free, which is great for starting out.

    

  • Paste the RSS feed URL you copied and click "Add" or "Save."

    

• Step 3: Configure Feed Settings:

  • This is an optional step because you may want to set preferences for how often you want to check for new updates.

  • You can set up filters to exclude certain keywords or topics you are not interested in, ensuring that you only receive content that is relevant to you.

    

3. Integrate RSS Feeds with Your Automation

Step 1: Set up a new automation scenario: For instance, you might create an automation that posts new articles to your social media accounts or adds them to a newsletter.

• Add the “RSS” module as the trigger. Choose option “Watch RSS Feeds item”.

  

• This option monitors the feed continuously and triggers an action as soon as a new item is published.

Step 2: Paste the URL and choose a maximum number of returned items:

• When adding an RSS feed, the URL is provided automatically. Just copy and paste this URL into the "URL" box. Ignore the “date from” and “date to” fields and set the number of items to retrieve as “5”.

  

Step 3: Test and Refine

• Run tests to ensure your automation works as expected. You can adjust settings and filters as necessary to improve accuracy and relevance.

• Here’s the result you’ll get:

  

4. Handling RSS Feed Data

Once you retrieve the RSS feed data, format it to meet your specific needs. This may include converting the data into a more readable format or extracting particular details from the feed.

1. Use the HTTP GET module: To get the full article content, add “HTTP” module to your scenario, and use the URL from the RSS feed to fetch the complete article.

   
   

2. Convert to Text: If the fetched content is in HTML, add the HTML to Text module after fetching the content. Choose Data from HTTP module. It will remove HTML tags and formatting, leaving you with plain text.

   
   

3. Store or Use Content: Once you have the plain text content, you have several options for storing or using it.

   

Here is the entire workflow:

Level 5: Automating Form Submissions to Welcome Emails

1. Introduction to automated email responses.

Automated email responses are digital tools that send pre-set replies based on certain conditions, acting like virtual assistants for routine email tasks. This system manages high volumes of emails, providing immediate responses without manual input, freeing up time for more complex tasks.

2. Step-by-Step Guide to automate in Make

Step 1: Build the front-end interface using Paperform

• Start by signing up for Paperform. You can choose a template or start from scratch with a blank page. But it’s better to build a form from the beginning to adjust it.

• For instance, you're launching a service named "Tweetify" that transforms YouTube URLs into viral tweets, here’s the form you may have:

  

Step 2: Connect Paperform submissions to Make

• The first step is to add a Paperform with new submission as the initial element and connect it with your previous form named “Tweetify”.

  

Step 3: Draft the Welcome Email:

• Add the Email module to your scenario.

• Create a connection to your email. Accounts ending with “gmail.com” may encounter some errors.

  

• It is recommended to use your Microsoft account.

  

• Draft your welcome email. Use placeholders to personalize it with the user's name and other details from the form.

  

Step 4: Test Your Automation:

• Submit a test form to ensure the automation works. You can check the Drafts section in your email to see the Welcome Email created by your automation.

  

Level 6. Automating Video Transcription and Summarization

In this level, we'll learn to automate YouTube video transcription and summarization by combining RSS feeds, Python, and GPT. This will streamline your workflow, letting you focus on more important tasks. By the end, you'll have an automated system that gathers video URLs, transcribes the content, creates summaries, and emails them to you.

1. Setting Up the Tools

• Required Tools: Make (or any other automation platform), ChatGPT (or any other AI language model).

2. Step-by-step guideline

1. Retrieve Video URLs with RSS Feeds

• Use an RSS feed to monitor your favorite YouTube channels. Create an RSS feed for a YouTube channel and get the feed URL.

  
  

• Add “RSS” module and choose option “Retrieve RSS feed items”. Paste the URL feed and choose 1 returned item to see.

  

2. Fetch the Transcript

• Add 4 modules in Level 3, including: HTTP - Make a request; Text parser - Match pattern; HTTP - Make a request; Text parser - HTML to text (excluding the 2 modules “Tools” and “Google docs”) to Make.

  

3. Use ChatGPT to summarize the transcript.

• Add the ChatGPT module to your scenario to summarize the transcript. Choose option “Create a completion”.

• You need to add a connection to ChatGPT using your secret API key.

  

• Access API keys - OpenAI API to retrieve your API key.

  

• Use a prompt to summarize the transcript text. For example, "Summarize the following YouTube video transcript in a concise and informative way."

  

4. Send the Summary via Email

• Add the Email module to your scenario which is similar to how we set up the Email module in level 5. You might want to revisit that setup to ensure everything is configured correctly.

  • Remember to fill in email of your recipients.

  

• Run Test: Trigger a test to ensure the transcription, summarization, and email sending processes work correctly.

• Verify Email: Check the recipient's inbox to confirm the summary email was sent and appears as expected.

Here is your entire workflow:

Level 7: Bundling Multiple Content Sources

1. Introduction to bundling RSS feeds

In Level 7, we'll cover how to merge content from multiple YouTube channels into a single feed. This feed will be used to transcribe and summarize the videos. Finally, the summaries will be compiled into one document. This approach helps you keep track of different sources efficiently and keeps everything organized in one place.

2. Create RSS Feeds from multiple channels

1. Create an RSS feed

   • Go to RSS.

   • For each YouTube channel you want to follow, enter the URL of the YouTube channel and RSS will generate an RSS feed for it.

   • Repeat this process for all the channels you want to monitor.

2. Bundle these feeds together

   • Go to the "Bundles" section in RSS and create new bundle.

     

   • Name the bundle and add all the individual YouTube channel RSS feeds you created.

     

   • This will give you one combined RSS feed URL. You can see the number of feeds in your overview. Now, you have the URL for a feed with multiple YouTube channels, prepare to work some magic in Make.

     

3. Automate Bundling in Make

Step 1: Fetch Video URLs from the Combined RSS Feed

• Create a new scenario in Make as usual.

• Add an RSS module to retrieve the combined feed for new video URLs. Enter the combined RSS feed URL and set the module to fetch the latest video URLs from the feed.

  

Step 2: Get the Transcript for Each Video

• As in level 3, add 4 modules, including: HTTP - Make a request; Text parser - Match pattern; HTTP - Make a request; Text parser - HTML to text (excluding the 2 modules “Tools” and “Google docs”) to Make. You might want to revisit that setup to ensure everything is set up right.

  

Step 3: Summarize Each Transcript: It's going to be the exact same process as we had before in Level 6.

• Use ChatGPT to summarize the transcript by adding the OpenAI module next to Text parser.

  

Step 4: Store Summaries in a Google Docs

• Add the Google Docs module to your scenario. Connect to your Google Docs account and create a new Google Document or select an existing one to compile the summaries.

  

• Configure the module to insert each summary as a new paragraph in the document.

Step 5: Run the scenario

• Before clicking on 'Run once,' remember to select where to start with the CSS module. Choose the most optimal option for your needs.

• And here is the result:

This way, you can save time and keep organized, ensure always have the key points from your favorite videos in one easy-to-read document. It’s like having a personal assistant who watches videos for you and gives you the highlights!

Level 8. Converting Text to Audio with ElevenLabs

1. Introduction to Text-to-Speech Automation

Text-to-speech (TTS) technology converts written text into spoken words, similar to having a robot read your text aloud. This is useful because it makes content more accessible, such as turning written articles into audio for people who prefer listening. It can also save time when you need to create audio content quickly.

2. Converting Text to Speech and Sending via Email

Step 1: Prepare the Summarized Text

• Make sure you have the summarized text ready in Level 8. Store this text in a Google Document or another easy-to-access place.

Step 2: Set Up ElevenLabs

• Sign up and log into ElevenLabs. Go to the settings and generate an API key. This key will allow you to connect ElevenLabs with other platforms.

  

Step 3: Integrate ElevenLabs into a Scenario in Make

• Actually you can duplicate the old scenario in Level 7 to generate another automation. Add the ElevenLabs module to your scenario. Enter the API key from ElevenLabs into the module.

  
  

• Configure the ElevenLabs module to convert the fetched text to speech. You can choose your favorite voice, model, and finally link the summarized text from the Google Docs module to the text input field.

Step 4: Send Audio via Email

• Add the Email module to your scenario. It’s not different from Email module you set in the previous level so just go ahead and set up it to fit your needs

• Additional suggestions:

  • Enter the recipient’s email address.

  • Set the email subject, such as "Your Summarized Content in Audio."

  • Write a brief message for the email body.

  • Attach the audio file created by ElevenLabs to the email.

Step 5: Test the Automation

• Run the scenario to check if the summarized text is converted to speech and sent via email correctly.

  

• Check the recipient's inbox to make sure the email with the audio attachment was received.

By following these steps, you’ll be able to easily convert text summaries into audio files and send them via email, making your content more accessible and engaging.

Level 9: Creating a Personal GPT Model

1. Introduction to training a GPT model on personal data

Training a GPT (Generative Pre-trained Transformer) on your data allows you to create a custom AI model that understands and responds based on your unique content. This can be incredibly useful for automating tasks like customer support, generating content, or providing personalized responses. In this guide, we will walk through the steps to collect your data, train a GPT model using it, and integrate the model into your workflow.

2. Step-by-step tutorials to build your own GPT

Step 1: Prepare Your Data

Don't worry if you don't have on-hand materials. You can gather video transcripts, blog posts, and articles from YouTube, RSS feeds, or other sources.

Remember, in Level 7, we set up the RSS, HTML and Text Parser modules to fetch transcripts from multiple sources. Now, continue from there and move on to the next steps I'll guide you through.

• Add another Google Docs module to gather all your notes and resources into a well-organized digital format.

  

Step 2: Access GPT-4 Playground

• Sign up on the OpenAI platform if you don’t already have an account.

• Navigate to the GPT-4 Playground on the OpenAI platform. Start a new project and select the option to create a custom GPT.

  

Step 3: Train your GPT

• Upload the compiled document as the training data.

• Configure settings like response length, tone, and style. Think of this step as setting the guidelines for how your assistant should behave.

• Instructions:

  • Provide clear instructions on how the GPT should use the uploaded data to generate responses. For example, you can instruct it to answer questions in your style, focusing on specific topics like AI automation.

  • Example: "Use the provided data to answer questions in the style of AI Fire, focusing on AI news topics”

    

Level 10: Developing a Unique Tone of Voice

1. Why should use a distinctive tone of voice?

Personalizing AI responses to match your speaking style makes interactions feel more genuine and engaging. This involves adjusting the AI's tone and phrasing to reflect how you naturally communicate.

2. Step-by-step guide to create a personalized voice

Step 1: Collect Your Data

• Gathering text from various sources that showcase your writing style. This can include personal blogs, articles, speech transcripts, or any written material you’ve created. For example, if you have written many blog posts, collect these as your data.

• Store all these texts in a digital format, such as Google Docs. This document will serve as the training data for the AI.

Step 2: Set Up on the OpenAI Platform

• If you didn’t generate your personal GPT in Level 9, go back and create one.

• Next, you'll train the AI to mimic your voice. Think of it as sharing your unique recipe with someone."

Step 3: Automate the process by using Make

• Basically, you can ask the AI sample questions and see how it responds in GPT. For example, test it with questions like “Explain the latest trends in AI automation.”

• But we’re here to make it more simple. Let’s create prompts for detailed voice customization. For example, we can add:

Create content that mirrors the distinctive style of AI Fire. The content should be:

Tone and Style: Professional yet approachable, maintaining a friendly and conversational tone. Avoid technical jargon and complex language; keep it simple and relatable.

Voice Characteristics: Engaging and clear, with a focus on delivering information in a straightforward manner. The content should be informative but also enjoyable to read, reflecting AI Fire's unique blend of authority and friendliness.

Audience Engagement: Tailored to a tech-savvy audience interested in AI and automation. Ensure the content resonates with their interests and provides value in an easy-to-understand way.

Content Focus: Highlight key points effectively. Ensure that the main ideas are clear and that any complex concepts are explained in a straightforward manner.

Structure and Length: Follow a structured approach with a clear introduction, body, and conclusion. The length should be appropriate for the content type (e.g., brief and punchy for social media, more detailed for blog posts).

For accuracy, refer to previous AI Fire content samples to ensure the new content aligns with the established style and tone

Here's the task:
"Explain in less than 100 words on how to build a personal GPT"

• Next, create a new scenario in Make and add OpenAI module, choose Message an assistant.

  

Step 4: Monitor and Refine

• Once testing is successful, activate your scenario in Make.

• Monitor the automation to ensure it functions correctly and make necessary adjustments based on feedback and results.

Creating a unique AI voice is like crafting a signature dish. By collecting and organizing data, setting up your AI model, and refining it through detailed analysis, you can make sure your AI-generated content sounds just like you.

Conclusion

Using AI can really boost how productive and efficient you are. It's like starting with a simple tool and slowly getting the hang of more advanced features. Start with easy AI tasks and work your way up to more complicated ones to get the most out of AI. 

Whether you're an individual or a business, AI automations help you accomplish more in less time. It's like having a smart assistant that learns and improves with use. By integrating AI into your daily routine, you’ll simplify your work and achieve your goals faster.

If you want to know more about AI and how it works, take a look at our other in-depth articles!

• Revolutionize Your Workflow: Make 1,000 Reels in Minutes with AI Automation!*

• Boost Your Study Game with These Top AI Tools for Students in 2024

• Master the Art of Creating an AI-Powered Service Desk FAQ Chatbot Easily*

• Claude vs ChatGPT: A Detailed Review of Which AI Assistant is Better and Why It Matters

• 20 AI Startups Raise over $1.1B (July 12, 2024 - July 19 2024)*