Make Your Own AI-Powered Research Assistant

Savvity
22 Mar 202419:16

TLDRThis video provides a step-by-step guide to creating a custom AI-powered research assistant that automates industry-specific research at a user-defined time and delivers the findings via email, Slack, or social media. The process involves using three key tools: Make, an automation platform for connecting various services; Perplexity AI, a search assistant; and the OpenAI Playground for custom GPT creation. The guide explains how to set up Make scenarios, purchase and install Perplexity's integration, and obtain API keys. It also covers creating prompts for Perplexity to ensure the desired research output and building actions within Make to execute the research and subsequent communication tasks. The video concludes with scheduling the research assistant and optionally formatting the research into a digestible format for different platforms using custom GPTs. A bundle including a JSON file with the flow blueprint and additional prompts is available for purchase.

Takeaways

  • 🤖 Utilize 'Make' to connect platforms without native integrations, like ChatGBT with Link D, Slack, or Medium.
  • 🧠 Integrate 'Perplexity AI' as a powerful search assistant to replace traditional search engines for comprehensive research.
  • 🚀 Access 'Chat GPT' through the Open AI playground to create custom GPT models tailored for specific tasks.
  • 🔗 Purchase the connection from Perplexity to Make, which is a new and valuable integration despite the cost.
  • 💡 Generate an API key from Perplexity and credit your account for the service to function.
  • 📝 Craft a detailed prompt for Perplexity to ensure the research output meets your needs.
  • 🔍 Test the prompt in Perplexity AI to verify the research results before proceeding with automation.
  • ✉️ Set up an automated email delivery for the research findings using Make, formatted in HTML for clarity.
  • ⏰ Schedule the research assistant to run at specific times, such as every Monday at 6:00 a.m., for regular updates.
  • 📅 Transform research into a digestible format using custom GPTs for various platforms like LinkedIn posts or newsletters.
  • 🔄 Use a router to direct the research output to different channels, such as email, Slack, or social media profiles.

Q & A

  • What is the main purpose of the AI-powered research assistant discussed in the video?

    -The main purpose of the AI-powered research assistant is to automate comprehensive research for a specific industry at a user-chosen time, and then deliver the findings in a well-formatted manner through email, Slack, or other communication channels, potentially even directly to social media profiles.

  • Which platforms are mentioned as key components for creating the AI-powered research assistant?

    -The key components mentioned are Make (an automation platform), Perplexity AI (an AI-powered search assistant), and access to the OpenAI playground for creating custom GPT models.

  • How does Make help in connecting different platforms for the research assistant?

    -Make helps by providing a no-code tool for automation that allows users to connect platforms that don't have native connections, such as connecting chatGBT with Link D, Slack, or Medium.

  • What is the role of Perplexity AI in the research assistant?

    -Perplexity AI serves as the core search engine for the research assistant, replacing traditional search methods like Google. It performs the actual research based on the user's prompt and returns the findings.

  • How does the user customize the research assistant to their specific needs?

    -The user customizes the research assistant by creating a specific prompt for Perplexity AI that defines the industry, field, and type of research required. This prompt is then tested and refined until it produces the desired research output.

  • What is the significance of defining the output as HTML in the research assistant's prompt?

    -Defining the output as HTML allows the research findings to be formatted nicely when sent via email or other communication channels that support HTML, making the information more readable and visually appealing.

  • How can the user schedule the research assistant to send updates at a specific time?

    -The user can set up a timing schedule within the Make platform, choosing the frequency (e.g., every Monday at 6 a.m.) for the research assistant to perform the search and send the findings.

  • What is the role of custom GPT models in transforming the research findings into a more digestible format?

    -Custom GPT models, created in the OpenAI playground, can take the raw research findings and transform them into specific formats, such as a LinkedIn post or a newsletter, making the content more suitable for different audiences and platforms.

  • How does the user ensure the research assistant's output is correctly formatted for different platforms?

    -The user can use custom GPT models that are specifically trained for different types of content (e.g., LinkedIn posts, newsletters) and provide clear instructions to these models on how to format the research findings.

  • What is the process for testing the research assistant's setup before finalizing it?

    -The user should run the entire flow, including the search and formatting actions, to test if the research is being handled correctly and the output is being sent as expected to the chosen communication channels.

  • Is there a bundled solution available for users who find the setup process overwhelming?

    -Yes, a bundled solution is available that includes a JSON file with the blueprint of the flow, a prompt engineer for creating complex research prompts, and other tools to simplify the setup process. This bundle can be purchased from a provided link.

  • How can users get support or request additional videos on creating AI-powered automations?

    -Users can follow the channel, like the video, and leave a comment below expressing their interest in more videos on creating AI-powered automations or requesting specific topics they'd like to see covered.

Outlines

00:00

🚀 Introduction to AI-Powered Search Assistant

This paragraph introduces the concept of an AI-empowered Search Assistant, a custom tool designed to automate research for a specific industry at a user-defined time. The assistant not only conducts comprehensive research but also formats and sends the findings via email, Slack, or direct posting to social media. The key components mentioned are Make (an automation tool), Perplexity (an AI search assistant), and access to the OpenAI playground for creating custom GPT models. The paragraph emphasizes the time-saving benefits of such a tool for those who conduct extensive research for work or study.

05:01

🔍 Setting Up the Research Assistant with Perplexity

The second paragraph delves into the setup process of the research assistant, focusing on configuring Perplexity AI. It outlines the importance of crafting the right prompt for Perplexity to ensure the research output meets the user's needs. The speaker provides a template for the prompt and encourages testing the output in Perplexity AI. The paragraph also guides on how to create actions within the Make platform, including setting up an email notification with HTML-formatted content for better presentation.

10:02

⏰ Scheduling and Customizing the Research Delivery

This paragraph explains how to schedule the research assistant to run at specific times, such as every Monday morning, ensuring that the user receives the research when needed. It also touches on the possibility of further customizing the research output into more digestible formats suitable for different platforms like LinkedIn posts or Slack messages. The integration of custom GPT models through the OpenAI playground is introduced to transform the research into specific content types, emphasizing the flexibility and power of the system.

15:03

📚 Finalizing the Research Assistant and Additional Resources

The final paragraph discusses the completion of the research assistant setup and offers additional resources to simplify the process. It mentions a bundle available for purchase that includes a JSON file to import the flow blueprint, a prompt engineer for crafting complex research prompts, and a copywriter prompt for creating digestible content. The speaker also provides a link to access the bundle and encourages viewers to subscribe and comment for more content on AI-powered automations.

Mindmap

Keywords

💡AI-Powered Research Assistant

An AI-Powered Research Assistant is an artificially intelligent system designed to autonomously conduct research on specific topics or industries at predefined times. In the video, it is portrayed as a tool that can significantly reduce the time spent on research by automatically gathering and formatting information, which it then delivers through various communication channels such as email or Slack.

💡Make (Zapier-like tool)

Make is a no-code automation tool that allows users to connect different platforms and services that do not have native integrations. It is pivotal in the video for creating a custom workflow where it connects services like Chat GPT with Slack and other platforms, enabling the automation of the research assistant's tasks.

💡Perplexity AI

Perplexity AI is an AI-powered search assistant that the video presenter has been using as a replacement for Google search. It is integral to the research assistant's functionality, as it is responsible for conducting the actual research based on the provided prompts and returning the findings.

💡Open AI Playground

The Open AI Playground is a platform where users can create and customize their own GPT models. In the context of the video, it is used to build custom GPTs that specialize in specific tasks, such as creating LinkedIn posts or newsletters from research data.

💡API Keys

API Keys are unique codes that allow developers to access and use external software or services. In the video, they are necessary to connect and authorize the use of Perplexity AI and Chat GPT within the Make automation workflow.

💡Scenario Creation

In the context of the Make tool, scenario creation refers to the process of setting up a sequence of automated actions. The video demonstrates how to create a scenario that involves using Perplexity AI to conduct research and then formatting and sending the results via email.

💡Prompt Engineering

Prompt engineering is the process of carefully crafting the inputs or 'prompts' given to AI systems to elicit the desired output. In the video, the presenter emphasizes the importance of creating the right prompt for Perplexity AI to ensure the research output meets specific needs.

💡Custom GPT

A Custom GPT, as discussed in the video, is a version of the GPT model that has been trained or fine-tuned for specific tasks. The presenter has created various custom GPTs for different purposes, such as writing SEO content or social media posts, which are then used to transform research into formatted outputs.

💡HTML Formatting

HTML (HyperText Markup Language) is used in the video to format the research output from Perplexity AI. The presenter chooses to have the research results in HTML format so that when the data is sent via email or other platforms, it maintains its structure and readability.

💡Scheduling

Scheduling in the automation context refers to setting up a timed trigger for the automation to run. In the video, the presenter schedules the research assistant to run every Monday at 6:00 a.m., ensuring regular delivery of research insights.

💡Router

A router in the video's context is a tool within the Make platform that directs the flow of data to different endpoints based on predefined conditions. It is used to send the research data to various channels, such as email, Slack, or directly to social media platforms.

Highlights

Create a custom AI-powered research assistant that automates industry-specific research at chosen times.

The research assistant can deliver findings via email, Slack, or direct posting to social media profiles.

Utilize Make.com to connect platforms without native integration, such as ChatGBT with Link D or Slack.

Perplexity AI replaces Google search for comprehensive research needs.

Access Custom GPT through the Open AI playground to create tailored language models.

Create an account on Make, Perplexity, and Open AI playground for building the research assistant.

Purchase the Perplexity to Make connection for streamlined integration.

Set up API keys and credit on Perplexity for accessing the AI search capabilities.

Craft a detailed prompt for Perplexity to ensure accurate and relevant research results.

Test the prompt in Perplexity AI to confirm the research output meets expectations.

Configure Make to create a chat completion action with Perplexity using the crafted prompt.

Automate the process to send research findings via email with HTML formatted content for readability.

Schedule the research assistant to run and send emails at specific times, like every Monday at 6:00 AM.

Use Chat GPT and custom GPTs to format research into digestible content for various platforms.

Create a custom GPT in the playground for specific tasks like writing LinkedIn posts or newsletters.

Send the formatted research to Slack channels, private Facebook groups, or directly to LinkedIn with a custom GPT.

Ensure character count is within platform limits, especially for LinkedIn posts.

The process can be overwhelming for beginners, but a bundle with a JSON file is available for easy setup.

The JSON file includes a blueprint of the flow, simplifying the process of setting up the research assistant.

Additional tools like the prompt engineer and copywriter prompts are included in the bundle for comprehensive assistance.

The AI-powered research assistant bundle is available at a low cost on the Stand store.