📊 A Detailed Guide to Transform Client Info into Insightful Reports with AI Research Automation

Introduction to AI Research Automation

Time is precious, and whether you're working in an AI automation agency, running a business, or just want to streamline your research, having an AI research assistant can be incredibly helpful.

Imagine having a team of robots doing all the heavy lifting for you. AI research automation dramatically reduces the time required to gather and analyze information. Instead of spending hours or even days on research, you can get detailed insights in minutes.

The magic of AI research automation happens when various tools work in harmony. The system operates through data collection from your customers, extraction, research and verification, insight generation, and report synthesis.

In this article, I'll show you how to create an AI-powered research assistant using three language models: Claude, Llama, and ChatGPT. This assistant will help you gather customer’s business’s information, check, and synthesize research, making your work easier and more efficient.

Now, let’s move on to setting up the tools needed for this powerful automation!!

I. Setting Up Your AI Automation Tools

First things first, let's get all the tools you'll need to set up your AI research assistant. We'll start by listing the necessary tools and why they are important for our AI automation setup.

1. Make

1. Purpose: Make is a no-code automation platform that connects various apps and services to automate workflows. Think of it as the glue that holds everything together.

   

2. Why It's Important: Make will be the backbone of your automation process. It connects different tools, like Paperform and Google Docs, and sets up workflows that run automatically.

We've provided a step-by-step guide in the AI Fire Beginner’s Tutorial before. If you encounter any issues, just refer to that guide, and you'll find all the answers you need.

2. Paperform

1. Purpose: Paperform is used to collect client information through customizable forms. Think of it as the digital clipboard that gathers all the details you need.

   

2. Why It's Important: Having a structured way to collect information ensures that you gather all necessary details right from the start. This makes the rest of the automation process smoother and more efficient.

3. Create an Account: If you don’t already have a Paperform account, sign up at Paperform. It's quick and easy, just click on “Try Paperform Free”

   

   • Fill Out all the Information: Let's add your interests, company name, or terms that accurately describe your business to help Paperform understand you better. This will ensure you get a more personalized and improved experience.

     

3. Perplexity

1. Purpose: Perplexity is an AI tool used to conduct in-depth research. It can gather and verify information from various sources. Think of it as your super-smart research assistant.

   

2. Why It's Important: Perplexity ensures that the data you use in your reports is accurate and comprehensive. It cross-references information from multiple reliable sources, acting as a fact-checker to provide you with dependable data.

3. Access Perplexity: 

   • Sign up: If you don’t already have an account, sign up at Perplexity. The sign-up process is quite straightforward. Just go with your Google account.

   

   • Get API Access: Once you have an account, navigate to the API section to get your API key. This key is the secret handshake that lets your systems talk to each other.

These tools are the building blocks of your AI research assistant. Each one plays a crucial role in ensuring the AI automation runs smoothly. Now that we have our tools, let's move on to creating the initial form where clients will input their information.

II. Creating the Initial Form for AI Automation

Next, we'll create a form to collect essential client information and book appointments. This form will be the starting point for our AI automation process. Let's break down the steps to create an effective form using Paperform.

1. Collecting Client Information

Step 1: Create a Form:

• Choose a Template: You can choose one of the pre-set templates to customize, which will save you a lot of time and effort. These templates are designed to make the process much easier so that you can focus on adding your personal touches.

  

• Start a New Form: When choosing to start with a blank template, you’ll be greeted by a blank canvas where you can add various fields to collect the necessary information.

Step 2: Add Form Fields:

• Include fields for Information: For example, you may want to collect the client's name, email, company name, website, biggest challenges, and dream outcome, don’t forget to add an appointment scheduling section. Here’s an example of what your form fields might look like:

  

• Add New Questions: You can also add your own questions to fit your needs. Just click on the "Add Another Question" button, type your question, and edit anything you want.

  

Step 3: Customize the Form Appearance:

• Themes and Styles: Go to the "Theme" section to select fonts, colors, and styles that match your brand. A well-designed form looks professional and encourages clients to complete it.

• Add a Logo: Including your company logo at the top of the form can enhance brand recognition and trust.

  

2. Configuring After-Submission Settings

Step 1: Create a Success Message: Set up a thank you message that appears after the form is submitted. This can be a simple "Thank you for your submission! We’ll get back to you soon."

Step 2: Test the Form: Before making the form live, perform a few test submissions to ensure everything works correctly.

• View the Form: Click on “View form” button on the right top corner to see your Paperfrom in reality. If you’re not happy with any field, come back to the dashboard and fix that.

  

• Test Submissions: You can also test the form by filling in random information to see if it runs correctly. This helps ensure everything works as expected before using it for real data.

  
  

Step 3: Make the Form Live

• Share the Form: Once you're satisfied with the form, make it live and share the link with your clients. You can embed the form on your website, share it via email, or provide the link directly.

  

Now that you have a form to collect client data, let's move on to extracting valuable information from the client's website.

III. Extracting Website Data for AI Research

In this section, we'll learn how to fetch and convert website data into readable text. This step ensures that our AI has the necessary information to work with. We’ll use Make to automate the process of extracting data from the client’s website.

1. Adding Paperform as a Trigger

First, we need to set up a trigger that starts the automation when a new form submission is received from Paperform.

• Step 1: Add Paperform Module: After creating a new scenario in Make, let’s add the Paperform module as the trigger.

  

• Step 2: Select Event: Choose "New Submission" as the event. This will trigger the scenario every time a client submits the form.

  

• Step 3: Connect Your Paperform Account: Follow the prompts to connect your Paperform account and select the form you created in the previous step.

  

• Step 4: Submit a Test Form: Submit a test form with random information to check the results and ensure everything runs smoothly.

  

3. Fetching Website Content Using HTTP Module

• Step 1: Add HTTP Module: Click the "+" button to add a new module to your scenario. Type "HTTP" in the search bar and select “Get a file” action.

  

• Step 2: Configure HTTP Module: Insert the URL of the client's website provided in the form submission. Use the dynamic field from the Paperform trigger to insert the URL.

  

• Step 3: Test the HTTP Request: Click on the "Run Once" button to test the HTTP request. Review the output of the HTTP request and you’ll see that it has fetched the HTML content of the client’s website correctly.

  

4. Converting HTML to Text Format

After fetching the HTML content from the client’s website, the next step is to convert this content into plain text. HTML content is full of tags and elements that make it hard to read and analyze directly. By converting it to text, we can easily extract and summarize the relevant information.

• Step 1: Add HTML to Text Module: One more time, just click the "+" button and add that module into the scenario.

• Step 2: Configure HTML to Text Module: Set the input to the content fetched by the HTTP module by choosing Data from the former module.

  

• Step 3: Test the HTML to Text Conversion: It’s time to see the magic happens, you’ll see that the module is correctly configured and can successfully convert the HTML content into plain text.

  

5. Summarizing Website Content

In this step, we’ll use ChatGPT to summarize the plain text content extracted from the website to get a clear and concise overview of the client’s business, which is essential for generating accurate and relevant insights.

Step 1: Add ChatGPT Module: Type "ChatGPT" in the search bar and select “Create a completion” action.

Step 2: Configure ChatGPT Module: One note here is to create a connection to OpenAI and run it, you need to have some credits which cost about $5, but it’s really worthy and reasonable to use.

• Model: Select the appropriate ChatGPT model (e.g., GPT 4o).

• Prompt Messages: Set up a prompt to instruct ChatGPT to summarize the website content. For your convenience, we’ve generated an optimized prompt here for you to copy:

You will be provided with the copy of a website. Your task is to carefully read and comprehend the content of the website. After analyzing the text, generate a concise summary of the key points and main ideas discussed in the website. The summary should be informative, helpful, and to the point, with a maximum length of 35 words.

When creating the summary, focus on the following aspects:

- Identify the core message or theme of the website.
- Highlight the most important facts, statistics, or examples that support the main points.
- If applicable, briefly mention any significant conclusions or recommendations made in the website.

Avoid using unnecessary fluff or filler phrases, and do not include introductory phrases like "this website is about." Instead, concentrate on directly communicating the essential information in a clear and efficient manner.

To ensure your summary is effective, consider the following tips:

- Use concise and precise language to convey the key ideas.
- Prioritize information based on its relevance to the main message.
- Ensure that the summary can be easily understood by someone who hasn't seen the website.
- Read through your summary to check for clarity, coherence, and adherence to the 35-word limit.

Remember, your goal is to provide a comprehensive yet succinct overview of the website's content, enabling readers to quickly grasp the core message and main takeaways. Do not repeat the title.

Then, write 5 core features in bullets with emojis like this:

* 📝 Summarize any website
* 🔍 Identify core message and theme
* 📊 Highlight key facts and statistics
* 💡 Mention significant conclusions
* ⚡ Concise 35-word summary

The website URL: {{3.data}}

• Input Text: Use the output from the HTML to Text module.

  

Step 3: Test the ChatGPT Summary: After Run the module, you can have a quick review of the output to ensure that it has successfully summarized the website content and listed the main features.

IV. Conducting In-Depth AI Research Using Perplexity

Now that you've summarized the website content, it's time to dig deeper. We'll use Perplexity, an AI tool that acts like a research detective, to gather more detailed and accurate information about the client's business.

1. Understanding the Role of Perplexity in AI Research

Perplexity helps you find valuable information and verify facts, ensuring your research is accurate and reliable. So, the format and the file I'm going to show you initially include Research A and Research B, each one has its own job.

Research A: Gather Interesting Facts

Research A’s job is to uncover 10 fascinating facts about the client’s business. Imagine Perplexity as your chief research officer, tirelessly exploring both the client’s website and other reliable external sources. The goal is to gather the most interesting and relevant information.

For example, you don’t want outdated information like what people thought about the wheel in the Stone Age. Instead, focus on recent and relevant facts from 2024 and beyond.

Here’s the Research A:

Your task is to find 10 intriguing facts about this business: {{4.data.fll64.value}}. Thoroughly explore the website and use reliable external sources to gather the most relevant information. Ensure each fact meets these criteria:

1. Well-Sourced: Provide the exact page, section, or article where the fact is found, using in-text citations or footnotes.

2. High-Quality: Verify accuracy by cross-checking with multiple reputable sources. Avoid outdated or incorrect information.

3. Up-to-Date: Use facts from the last 12 months. Clearly indicate the time period for any historical data.

4. Diverse Coverage: Explore various topics and sections of the website for a well-rounded understanding. Avoid focusing on a single area.

5. Relevance: Focus on significant and valuable facts that clarify the website's main themes, objectives, or unique features. Avoid trivial information.

6. Clear Documentation: Present each fact clearly and concisely, with brief explanations if needed. Use bullet points or numbered lists for readability.

7. Avoid Redundancy: Ensure each fact is unique and adds new information. Avoid repeating similar facts.

8. Data-Driven Insights: Include interesting statistics or data points, with citations.

9. Business Relevance: Focus on facts that provide insights into the company's operations, target market, competitive landscape, or growth opportunities to support a well-informed consultation.

10. Thought-Provoking: Include at least one or two surprising or lesser-known facts that spark curiosity and show deep understanding of the company.

Present the facts in a numbered list, with each fact followed by its specific source(s) using in-text citations or footnotes. Include sources as naked URLs.

Research B: Fact-Check Information

Research B’s job is to fact-check the information gathered by Research A. This step is crucial to ensure the accuracy and reliability of the facts. Think of it as a quality control process, making sure Research A is telling us the right stuff. Research B verifies each fact and provides reliable sources for each piece of information.

Here’s the Research B:

Please analyze the following output from the previous model:

{{6.choices[].message.content}}

For each bullet point in the output, perform the following tasks:

For each bullet point in the output, follow these steps:

1. Verify if the bullet point has a source URL.

- If it lacks a source URL, find a reliable source to support the information.
- If a source URL exists, include it with the bullet point.
- Ensure to use the naked URL for every source.
- If no reliable source can be found, remove the bullet point.

2. Fact-check each bullet point.

- Confirm the accuracy of the information using reliable sources and fact-checking methods.
- If any information is inaccurate, correct it or remove the bullet point if it cannot be corrected.

3. Retain the original format of the output, including emoji bullet points and headers.

4. For each bullet point that passes fact-checking, include the source URL(s) used for verification.

5. Provide only the updated list without any additional information or introduction, such as "sure, here's the updated list you asked for." Do not add any extra text before or after the 10 facts.

Ensure to use the naked URL for every source.

Please provide the updated output with fact-checked information, source URLs, and any necessary corrections, while keeping the emoji bullet points and headers intact.

2. Gathering Facts about Client’s Website in Research A

• Step 1: Add the First Perplexity Module: As usual, add Perplexity module next to OpenAI and select “Complete a Chat completion” in it. This action will generate a response based on the input you provide, guiding Perplexity to perform the research tasks you need.

• Step 2: Configure Perplexity Module:

  • Create a connection: Enter your Perplexity API key. If you don't have an API key yet, let’s visit Perplexity again and obtain it from your account settings.

  

  • Choose the Model: You can opt for the LLaMA-3-70B-Instruct model. This model has the ability to handle complex instructions, process large amounts of data efficiently, amaking it ideal for advanced AI tasks.

  • Messages: In the content box, paste the Research A prompt I provided here and replace "" with the URL link in Paperform. And the Role is gonna be for User.

  

• Step 3: Test the Module: Click on the "Run Once" button to test the Perplexity research configuration. Remember to submit another Paperform whenever you want to Run the whole automation.

  

Wonderful, right? You can now see 10 interesting facts about AIQrArt, the website I linked in Paperform, each accompanied by a clear and detailed description. This makes it easy to understand the unique aspects and features of AIQrArt, giving you a comprehensive overview of what the site offers.

3. Fact-Checking Information in Research B

• Step 1: Add the Second Perplexity module: Right-click the Research A module and then choose "Clone." This means you are duplicating the module, so you don't have to set it up from the beginning again.

  

• Step 2: Customize Research B module: Now, you need to change the role of this module to match Research B's job. To do this, fill out the message again using the prompt I provided earlier. Then, replace the placeholders "" with the content from the Research A module.

  

• Step 3: Run and Test the Configuration: Click on Run Once again with another submission form and see the magic!

  

That's incredible! The quality of the 10 facts from Research A has been upgraded to the next level with detailed descriptions and clear explanations. This enhancement makes the information much more insightful and valuable.

V. Combining Insights from AI Models

After gathering detailed information and verifying facts with Perplexity, it's time to combine insights from multiple AI models such as ChatGPT and Llama to create a comprehensive and actionable report.

You might wonder why need need to use various AI Models. So, combining insights from different AI models ensures a more holistic view of the client's business. Each model has its strengths and can provide unique perspectives, making the final report more comprehensive and reliable. Let’s dive in!

1. Adding ChatGPT as the First AI Language Model

Basically, it's the same process as before. To keep it simple, I'll provide you with the new message to copy and paste into the Message box. The other placeholders match those in the ChatGPT module we added earlier.

Here’s the message:

As an AI-powered Opportunities Bot, your task is to find valuable insights and spot areas where our AI automation agency can best serve our prospective clients. To do this, analyze three key pieces of information:

Details about our AI automation agency, including our services, expertise, and successful case studies.
Information about the prospective client, such as their industry, size, current processes, and technology stack.
The client's goals, aspirations, and biggest challenges or pain points.
Using this data, create three engaging paragraphs that highlight specific opportunities where our agency can provide the most value to the client. Emphasize how our services can help them overcome their challenges, optimize their processes, and achieve their goals.

In the first paragraph, address the client's most pressing pain point and explain how our agency's relevant services can solve this issue directly. Provide a clear and concise solution that shows our understanding of their needs and our ability to deliver results.

In the second paragraph, focus on the client's dreams and aspirations, and showcase how our agency's expertise can help them achieve these goals. Highlight the potential benefits of our services, such as increased efficiency, cost savings, and improved customer satisfaction.

In the third paragraph, identify a unique opportunity that the client may not have considered before. Demonstrate our agency's innovative thinking and proactive approach to problem-solving. Explain how this opportunity aligns with their goals and how our services can help them seize it.

Finally, conclude your analysis with three thought-provoking questions to ask the client during our meeting. These questions should spark meaningful discussions, uncover additional insights, and show our genuine interest in their success. Number these questions as follows:

[Question 1] 🤔
[Question 2] 💎
[Question 3] 💷
A good question helps facilitate positive discussions towards understanding the client's business problems or dream outcomes. For example:

You mentioned you are struggling with client acquisition, which channels have you explored so far? 🤔
I noticed you offer digital marketing services to design agencies on your website, but you wanted to focus on a different product. Is that something you would like to explore? 💎
To help you achieve your goal of scaling effectively using AI, please share your client acquisition cost and lifetime customer value 💷
Tailor your language and tone to the specific client and their industry. Use clear, concise, and persuasive language that showcases our agency's value proposition and sets us apart from competitors. Your insights will be key in helping me prepare for a productive and impactful meeting with the prospective client.

Here is the data:

Client’s dream outcome & problems:
Data & research about the client:
Information about my business:"

Now it's time to see if you've improved in using Make. Choose the content and fill in all the information about the client's dream outcome and problems, Data & research about the client.

Don't worry if you can't find the placeholders to fill in; here's an image to guide you. Follow it step-by-step to ensure everything is filled out correctly. About information about your business, let’s prepare your own file such as Google docs to paste here.

2. Adding Llama as the Second AI Language Model

This step is quite straightforward for you. First, add the Perplexity module right after the ChatGPT module in your workflow. Then, configure the Perplexity module in the same way you did for the ChatGPT module.

Let’s use the same message and role settings to ensure consistency. This will help maintain a smooth and effective process throughout your automation.

Here you can step further by submit a form and test the whole automation. Don’t be shocked by the results because of its amazing content. Let’s see the result you might receive:

VI. Synthesizing AI Research Information

1. Understanding the Role of Synthesizing in AI Research:

After generating insights from each AI model, we now need to synthesize the information to create a cohesive final output. This synthesis involves combining the strengths of each AI model to create a detailed and actionable summary for your client's business.

Synthesizing information from various sources and models ensures a well-rounded and accurate view of the client's business. It integrates diverse insights into a unified report, making it easier for decision-makers to understand and act upon the findings.

2. Using “King” Prompt to Get Started

Here, we'll use the OpenAI module again but with a totally different message to merge responses from these two AI models, creating a single, polished report. I called this new prompt the “King”.

Because we used OpenAI several times in this automation, let's simply clone it and edit the message. It sounds like a piece of cake, right? This will save time and ensure consistency throughout the process.

Finally, this is the King prompt for you to instruct ChatGPT to synthesize the insights from ChatGPT, and Llama.

Your mission is to analyze the responses from two AI models, each addressing the Opportunities Bot Prompt. Follow these steps to create an enhanced final output:

1. Carefully review the two responses, focusing on their key insights, recommendations, and questions.

2. Compare the responses to identify any hallucinations, inaccuracies, or inconsistencies. Highlight the strengths of each response, such as unique perspectives, compelling arguments, or well-crafted questions.

3. Merge the best parts of each response to create a cohesive, persuasive, and insightful final output. Leverage the strengths of each model to enhance the overall quality and impact of the final response.

4. Ensure that the final output maintains the specified format:
   - Three compelling paragraphs highlighting specific opportunities where our agency can provide the most value to the client.
   - Each paragraph should focus on a different aspect, such as addressing pain points, realizing dreams and aspirations, or identifying unique opportunities.
   - Conclude with three thought-provoking questions designed to spark meaningful discussions and demonstrate our genuine interest in the client's success.
   - Number the questions as follows:
     1. [Question 1] 
     2. [Question 2] 
     3. [Question 3] 
   - Include a relevant emoji after each question.

5. Present the final output as a standalone response, without any additional commentary or explanations.

6. do not write this as a letter. keep each paragraph to a maximum of 75 words and write this as a succinct, informative executive report. You must add three questions at the end of the analysis:

paragraph 1
paragraph 2
paragraph 3
question 1
question 2
question 3

Remember to maintain a clear, concise, and persuasive tone that showcases our agency's value proposition and differentiates us from competitors. The final output should be a powerful tool to help prepare for a productive and impactful meeting with the prospective client.

Do not copy just one of these. You must produce an executive report combining elements from all of them. So, to confirm, you will have three paragraphs of analysis followed by three questions. This is an executive report.

You must learn at least one thing from each response.

Output 1:
Output 2: 

For the last 2 lines in this prompt, let’s add output 1 and output 2 from 2 previous models. With a synthesized report ready, it's time to generate the final document in Google Docs.

VII. Generating AI Research Reports in Google Docs

Once you've synthesized the information from various AI models, the next step is to generate a comprehensive report in Google Docs. This report will present all the insights in a clear and structured format, making it easy for stakeholders to review and act on the findings.

1. Designing the Report Template

Designing the Google Docs template is a crucial step in generating comprehensive and professional AI research reports. Your template will serve as the blueprint for the report, ensuring consistency and clarity in how the information is presented.

• Step 1: Create a New Google Doc: Go to Google Docs and click on "Blank" to create a new document.

  

• Step 2: Set Up the Document Layout: At the top of your document, add a placeholder for the report title. Divide the document into sections to organize the information clearly. Each section should have a heading and a placeholder for dynamic content.

  

• Step 3: Customize the Template: Choose fonts and styles that are professional and easy to read. Ensure proper spacing and alignment for readability. If applicable, add branding elements such as logos or company colors to the template.

  Here’s a simple structure you can use:

Your meeting
Who ��
{{CLIENTNAME}}
When ��
{{MEETINGDATE}}

Website ��
{{WEBSITE}}

Dream & outcome ��
{{DREAMOUTCOME}}

Website Overview
{{WEBSITEOVERVIEW}}

Opportunities King �

{{OPPORTUNITIESKING}}

ChatGPT �
{{CHATGPT}}

Llama �
{{LLAMA}}

2. Automating Document Creation in Make

• Step 1: Add the Google Docs Module: To generate documents automatically, let’s add the Google Docs module to your scenario and choose “Create a document from a template”

  

• Step 2: Connect Your Google Account: Follow the prompts to connect your Google account to Make.

• Step 3: Configure the Module:

  • Document ID: Choose the Google Docs template you created earlier.

  • Configure Data Flow Between Modules: The output from Perplexity, ChatGPT, Claude, and Llama should be mapped correctly to the Google Docs module.

  

• Step 4: Test the Scenario: Check the generated Google Docs report to ensure all placeholders are correctly populated and the information is accurate.

  

By following these steps, you can automate the creation of detailed and professional AI research reports in Google Docs. Once the document is generated, let's review and use it effectively for client meetings.

VIII. Reviewing and Using AI Research Reports

After automating the creation of your AI research reports in Google Docs, the final step is to review these reports and effectively use the insights to drive decision-making and strategy.

1. Run the whole automation:

   • Initiate the entire automation workflow in Make. Ensure all modules are properly configured and the automation runs smoothly from start to finish.

2. Open the Generated Report:

   • Open the Google Docs report generated by your automated scenario. You could find it in the designated Google Drive folder you specified.

   

3. Check for Completeness:

   • Ensure all sections of the report are populated with relevant data. Verify that all placeholders have been replaced with actual content.

4. Verify Accuracy:

   • Cross-check key data points with the original sources to ensure accuracy. This includes verifying facts, figures, and any strategic recommendations provided by the AI models.

5. Assess Readability:

   • Ensure the report is easy to read and understand. Look for any jargon or complex language that might need simplification. The report should be clear and accessible to all stakeholders.

Conclusion

Using AI to create and review research reports can make a huge difference in how you run your business. By following these steps, you’ll streamline your operations, make better decisions, and set your company up for growth.

Imagine you’re baking a cake. You wouldn’t mix the ingredients, bake, and decorate all by yourself if you had a top-notch kitchen team to help, right? Using AI for your research is like having that dream team in your business kitchen.

By letting AI handle the heavy lifting, you can focus on what you do best—making strategic decisions and growing your business. This approach not only saves time but also ensures you’re working with the most accurate and up-to-date information.

So, embrace these AI tools. They’ll help you stay ahead of the competition, run your business more efficiently, and achieve the success you’re aiming for!!!

If you are interested in other topics and how AI is transforming different aspects of our lives, or even in making money using AI with more detailed, step-by-step guidance, you can find our other articles here:

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

• Discover the Power of AI: A Comprehensive Beginner's Guide to Workflow Automation

• A Comprehensive Guide to Turn Your Inbound Emails into Profits with Smart AI Automation in Minutes

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

• Transform Your Daily News into a Personalized AI Podcast: Guide to Creating Engaging Audio Summaries*

*indicates a premium content, if any