Automate LinkedIn Outreach With AI Videos (Steal This System)

LinkedIn Outreach Automation Workflow: Summary

Key Concepts:

• LinkedIn Automation
• Web Scraping (Apollo, LinkedIn)
• Lead Generation
• Personalized Messaging
• AI-Powered Script Generation (OpenAI)
• Video Personalization (HeyGen, Creatomate)
• Automated Messaging (Phantom Buster)
• N8N (Automation Platform)
• Apify (Web Scraping Platform)

1. Overview of the Automated Workflow

The video details a fully automated LinkedIn outreach workflow designed to generate leads, personalize messages, and send connection requests and follow-ups, all on autopilot. The core idea is to leverage various platforms to scrape data, create personalized content using AI, and automate the sending of messages on LinkedIn.

2. Workflow Breakdown: Step-by-Step

The workflow is built using N8N, an automation platform, and involves the following steps:

1. Trigger: The workflow is triggered daily at 9:00 AM.
2. Web Scraping Leads from Apollo:
   • Uses Apify to scrape leads from Apollo.io, a lead generation platform.
   • Employs the "Apollo Scraper" actor in Apify (cost: $0.08 per 1,000 leads).
   • Scrapes approximately 500 prospects based on specified criteria (e.g., job title, location) defined in the Apollo URL.
   • Requires an Apify API key for authorization.
   • The URL used for scraping mimics Apollo's search filters (e.g., job titles like "CEO," location like "United States").
3. Filtering Leads:
   • Filters out leads without an email address or LinkedIn profile to ensure contactability.
4. Limiting Outreach:
   • Implements a limit (e.g., 10 per day) to stay within LinkedIn's messaging limits (ideally under 100 messages per week).
5. Enriching Data from LinkedIn Profiles:
   • Uses Apify to scrape individual LinkedIn profiles to gather data like name, title, job history, and other relevant information.
   • Requires an Apify API key.
   • This data is used to personalize messages and video scripts.
6. Data Transformation with Code Block:
   • Uses a JavaScript code block (generated by ChatGPT) to transform the JSON data from the LinkedIn profile into a single text file.
   • This simplifies the process of passing the data to OpenAI for script generation.
   • The code block handles variations in profile data (e.g., different numbers of education entries).
7. Filtering for Active LinkedIn Profiles:
   • Filters out profiles that return a blank value for the full name field, ensuring an active profile.
8. AI-Powered Script Generation (HeyGen):
   • Uses OpenAI (GPT-4 or a similar model) to generate personalized video scripts and LinkedIn messages.
   • Requires an OpenAI API key.
   • Employs a system of "system message," "user message," and "assistant message" to guide the AI:
     • System Message: Provides context (e.g., "Write me a script that is going to be turned into a video").
     • User Message: Provides input (e.g., the person's LinkedIn profile data).
     • Assistant Message: Defines the output format (e.g., "Create us a script").
   • The system message includes guidelines for the script (e.g., conversational tone, focus on the prospect).
   • Generates three types of messages: video script, initial LinkedIn connection message, and follow-up message.
   • Standardizes the output as JSON data with a "message" key for consistent referencing.
9. HeyGen Video Creation:
   • Uses HeyGen to create a personalized video using the generated script and a custom avatar.
   • Requires a HeyGen API key, avatar ID, and voice ID.
   • Avatar and voice clones are created by uploading a video of the user speaking.
   • Waits 2 minutes for the video to be generated and then retrieves the publicly accessible link.
10. Website Screen Recording:
    • Uses Apify to create a screen recording of the prospect's website.
    • Employs either the "Apify Screenshot Generator" or the "Ultimate Screenshot" actor.
    • Requires an Apify API key.
    • Configures settings like output format (video file) and maximum scroll duration (e.g., 10 seconds).
11. Video Rendering (Creatomate):
    • Uses Creatomate to combine the HeyGen video and the website screen recording into a final personalized video.
    • Requires a Creatomate API key and a pre-designed video template.
    • The video template is created in Creatomate using JSON data provided in the video.
    • The Creatomate node in N8N uses the Creatomate API to render the video.
12. Data Storage (Google Sheets):
    • Stores all generated data (prospect information, LinkedIn profile, messages, video links) in a Google Sheet.
    • The Google Sheet is used as a data source for Phantom Buster.
13. Automated Messaging (Phantom Buster):
    • Uses Phantom Buster to automatically send connection requests and follow-up messages on LinkedIn.
    • Requires a paid Phantom Buster account.
    • Connects to the Google Sheet to retrieve prospect data and messages.
    • Uses LinkedIn session cookies (obtained via the Phantom Buster Chrome extension) to log in and send messages.
    • The Google Sheet URL and the column containing LinkedIn profile links must be specified in Phantom Buster.
    • The invitation message and follow-up messages are pulled from the corresponding columns in the Google Sheet.
    • Daily invitation limits should be set conservatively (e.g., 20 per day) to avoid account restrictions.
    • Column headers in the Google Sheet must not contain spaces (use underscores instead).

3. Key Arguments and Perspectives

• Automation Efficiency: The workflow significantly reduces the time and effort required for LinkedIn prospecting and outreach.
• Personalization: AI-powered script generation and personalized videos increase engagement and connection rates.
• Scalability: The automated system allows for reaching a large number of prospects efficiently.
• Compliance: It's crucial to adhere to LinkedIn's terms of service and avoid sending excessive messages to prevent account restrictions.

4. Technical Terms and Concepts

• Web Scraping: Extracting data from websites programmatically.
• API (Application Programming Interface): A set of rules and specifications that software programs can follow to communicate with each other.
• JSON (JavaScript Object Notation): A lightweight data-interchange format.
• Actor (Apify): Pre-built web scraping tools within the Apify platform.
• Bearer Token: An authentication token used to grant access to an API.
• Rate Limiting: A mechanism to limit the number of requests a user can make to an API within a given time period.

5. Notable Quotes

• "The coolest part is is that 100% of this, like literally every single piece of this is going to be automated."
• "I didn't even write this myself. I just got chatbt to do this for me. Pretty much all the code that I use inside N&N now I just get chatbt to write."
• "Everyone likes to receive compliments. So, the more like if you put that in there, the chance I believe of connecting with somebody obviously increases, right?"

6. Synthesis/Conclusion

The automated LinkedIn outreach workflow presented in the video offers a powerful solution for lead generation and personalized engagement. By combining web scraping, AI-powered content creation, and automation tools, users can significantly streamline their prospecting efforts. However, it's essential to use the system responsibly, adhering to LinkedIn's guidelines and prioritizing personalization to build genuine connections. The speaker also emphasizes the importance of continuous learning and adaptation, encouraging viewers to explore the broader possibilities of automation in their businesses.