About this video

• Video Title: AI AUTOMATION BOOTCAMP DAY 3
• Channel: Jokotoye Emmanuel | Automation King
• Speakers: Jokotoye Emmanuel
• Duration: 02:08:00

Overview

This video is the third day of an AI Automation Bootcamp. The instructor, Jokotoye Emmanuel, covers essential topics such as importing and exporting workflows in N8, understanding execution logs for debugging, and the various types of nodes used in automation (trigger, action, logic). He also discusses the differences between cloud-based and local installations of N8, emphasizing the importance of building in public and practical application.

Key takeaways

• N8 Workflow Import/Export: Learn how to import and export N8 workflows to reuse, back up, share, and even sell them as digital products.
• Execution Logs: Understand the importance of execution logs for monitoring, debugging, and troubleshooting automation workflows, identifying successes (green) and failures (red).
• Node Types: Grasp the three main families of nodes: trigger (starts the workflow, like a doorbell), action (performs tasks), and logic (makes decisions, like IF or SWITCH statements).
• Local vs. Cloud N8: Differentiate between using the cloud version of N8 (easy, no setup) and local installation (requires Node.js, acts as a playground, but automations stop when the computer is off).
• Building in Public: Emphasizes the necessity of sharing your work and progress online (e.g., on LinkedIn, X) to build social proof and attract clients, even for introverts.

This video practically explains several key aspects of AI automation using the N8 platform:

• Importing and Exporting Workflows: The instructor demonstrates how to import and export workflows within N8. This practical skill allows users to reuse existing automation, back up their work, share creations with others, and even sell workflows as digital products. It's shown how to do this through the three-dotted menu on the dashboard, using options like "import from file" and "download."

• Execution Log for Debugging: A practical demonstration of how to use N8's execution logs is provided. The speaker walks through identifying successful (green) and failed (red) workflow runs, and how to click into specific executions to see detailed error messages and understand exactly where a workflow went wrong. This is crucial for troubleshooting issues like missing API keys or incorrect data formats.

• Node Functionality: The video practically explains the function of various nodes by categorizing them into trigger, action, and logic nodes. Examples like the webhook node (receiving data), cron node (scheduling tasks), and if/switch nodes (decision making) are explained through relatable analogies and use cases (e.g., a doorbell for a trigger, an alarm clock for a cron node).

• Local N8 Installation and Usage: The instructor provides a practical walkthrough of installing and running N8 locally on a computer using Node.js. This includes searching for Node.js command prompts, running the N8 installation command, and accessing the local N8 dashboard via a localhost link in the browser. The limitations of local installations (automations stop when the PC is off) are also practically highlighted.

• Building a LinkedIn Research Agent: A live demonstration of building and executing a workflow is showcased. The instructor uses a form trigger to input a LinkedIn URL and then demonstrates how the workflow scrapes the profile to generate a summary report, identifying pain points and suggesting solutions. This practical example illustrates how to use N8 for lead generation and sales intelligence.

The instructor practically demonstrated the building of the research agent by following these steps:

1. Workflow Design (Implicit): While the exact creation of the workflow wasn't shown in minute detail, it's implied that the agent was built using various N8 nodes. Key nodes likely used include:

   • A Trigger Node: Specifically, a Form Trigger was used. This means the workflow starts when a user submits data through a form.
   • An Input for LinkedIn URL: The form was designed to accept a LinkedIn profile URL.
   • Web Scraping Nodes: Nodes (likely involving HTTP requests or specialized scraping tools) were used to access and extract data from the provided LinkedIn URL.
   • Data Processing/AI Nodes: Nodes were used to analyze the scraped data, identify pain points, and generate summary reports. This likely involved API calls to AI models (like OpenAI, as mentioned elsewhere) for analysis and text generation.
   • An Output Node: This node was configured to present the summarized information, including pain points and potential solutions, in a readable format. It's implied that this output was then displayed or sent, possibly via email or to a dashboard.

2. Execution of the Workflow:

   • Accessing the Form: The instructor accessed the form associated with the research agent workflow.
   • Inputting Data: He pasted a LinkedIn URL (specifically Mr. Shum's profile) into the form.
   • Submitting the Form: Clicking the "Submit" button initiated the workflow.

3. Demonstrating Workflow Functionality:

   • Workflow Running Indicator: The interface showed that the workflow was active ("currently running").
   • Scraping Process: The video indicated that the agent was actively scraping the LinkedIn profile.
   • Displaying Results: After the workflow completed, the instructor showed the output: a summary report of the individual's profile, including their role, experience, company, interests, unique facts, similarities, and crucially, identified pain points and suggested AI-driven solutions. He used this output to explain how it could be used for pitching to clients.

Essentially, the demonstration focused on the triggering of the agent (via form submission) and the output it produced, which showcased its practical application in generating sales insights.

The video elaborates on the workflow process for the LinkedIn research agent by explaining the function of its components and then demonstrating its execution.

Workflow Components and Their Functions:

1. Trigger Node (Form Trigger):

   • What it is: This is the starting point of the workflow. It's activated when new data is submitted through a form.
   • How it was done: The instructor showed a form designed specifically for this agent. Users are expected to input a LinkedIn URL into this form. When the "Submit" button is clicked, this action triggers the entire automation. This is analogous to a doorbell; it signals that a task needs to begin.

2. Data Input (LinkedIn URL):

   • What it is: The specific piece of information the user provides to start the process.
   • How it was done: The instructor pasted a LinkedIn profile URL into the form. This URL is the essential input that tells the agent whose profile to research.

3. Web Scraping/Data Extraction:

   • What it is: This is the core process of accessing the LinkedIn profile and gathering relevant information. Since N8 might not have a direct, built-in node for scraping every website, the instructor likely used the HTTP Request node.
   • How it was done: The HTTP Request node would have been configured to:
     • Send a request to the provided LinkedIn URL.
     • Receive the HTML content of the LinkedIn profile page.
     • Parse this content to extract specific data points like name, job title, experience, company, interests, etc. This part often involves custom logic or integration with specialized scraping tools/APIs if LinkedIn's structure is complex or protected. The instructor mentioned that N8 doesn't have a built-in node for scraping every platform, implying the HTTP Request node is versatile for such tasks.

4. Data Processing and Analysis (AI Integration):

   • What it is: Once the raw data is scraped, it needs to be analyzed to identify key insights like pain points and potential solutions. The instructor implied the use of AI for this.
   • How it was done: It's highly probable that after extracting the data, it was fed into an AI model (like OpenAI, which the instructor later mentions using and recommends). The AI would be prompted to:
     • Summarize the individual's professional background.
     • Identify potential challenges or "pain points" based on their stated experience, industry, or roles.
     • Suggest relevant AI-driven solutions that could address these pain points.
     • The "Set Node" might have been used here to structure or format the data before sending it to the AI, or to organize the AI's output.

5. Output/Reporting:

   • What it is: Presenting the analyzed information in a clear and actionable format.
   • How it was done: The workflow was designed to compile all the gathered and analyzed information into a summary report. This report included sections like:
     • Personal Summary (based on their title and experience).
     • Pain Points (identified challenges).
     • Solutions (suggested AI-driven services or products).
     • Information about their company and interests. The instructor displayed this report on the screen after the workflow ran, demonstrating the tangible outcome of the agent.

Demonstration of Execution:

• The instructor navigated to the workflow's interface, which presented the form trigger.
• He then pasted a specific LinkedIn URL into the designated field.
• Upon clicking "Submit," the workflow visually indicated it was running.
• The instructor commented that the agent was scraping the profile.
• Finally, the generated report was displayed, showcasing the extracted and analyzed data, which the instructor then used to explain how someone could leverage this information for sales pitches.

In essence, the workflow starts with a user providing a LinkedIn URL via a form, then uses tools like HTTP requests (and likely AI integration) to scrape and analyze the profile, and finally presents a structured report of insights that can be used for sales or research purposes.

Yes, the form shown in the demonstration was indeed connected to n8n. The instructor explained and demonstrated this connection implicitly through the workflow's execution.

Here's how it works and how it was likely set up:

1. The Form as a Trigger: In n8n, forms are commonly used as trigger nodes. When you create a workflow that starts with a form, n8n provides you with a unique URL for that form.
2. Form Submission: Anyone who has access to that form URL can fill it out and submit data.
3. n8n Receives Data: When the form is submitted, n8n automatically receives the data entered by the user. This submission is what activates the trigger node and starts the rest of the workflow.
4. Workflow Execution: The submitted data (in this case, the LinkedIn URL) is then passed on to the subsequent nodes in the workflow, allowing the automation to proceed with scraping and analysis.

How it's done in n8n (General Process):

• Adding the Form Trigger Node: The instructor would have added a node from the "Trigger" category, specifically the "Form" node.
• Configuring the Form: Within the Form node's settings, fields are defined to collect the necessary information. In this case, a field for "LinkedIn URL" was created.
• Getting the Form URL: Once configured, n8n provides a unique URL for this specific form. This URL is what the user (or anyone else) would access to fill out the form.
• Connecting to Subsequent Nodes: After the Form node, other nodes (like HTTP Request, AI nodes, Set nodes, etc.) would be connected. These nodes receive the data submitted through the form as their input.

During the demonstration, when the instructor pasted the LinkedIn URL into the form and clicked "Submit," he was interacting directly with this n8n-generated form. The act of submission triggered the rest of the automation chain he had built within n8n.