Linkedin Outreach AI Agent

What This Agent Does

This intelligent automation agent transforms your LinkedIn outreach from a time-consuming manual process into a seamless, personalized connection machine. Every minute, it scans your Google Sheets prospect list, retrieves LinkedIn profile data, generates AI-powered personalized connection messages using Claude AI, and automatically sends connection requests—all while meticulously tracking progress and updating your spreadsheet in real-time.

Key benefits include:

• Save 15-20 hours per week on manual LinkedIn outreach tasks
• Increase connection acceptance rates by 40-60% through AI-personalized messaging
• Scale your networking efforts to reach 50-100+ prospects daily without lifting a finger
• Maintain complete visibility with automatic status tracking and message logging
• Eliminate human error with consistent, professional outreach every time

Perfect use cases:

• Sales teams building prospect pipelines
• Recruiters connecting with potential candidates
• Business development professionals expanding their network
• Marketing teams conducting influencer outreach
• Entrepreneurs growing their professional connections

Who Is It For

This automation is designed for growth-focused professionals who understand that networking is essential but don't have hours to spend on repetitive outreach tasks. Whether you're a sales development representative managing hundreds of prospects, a recruiter building talent pipelines, or a founder establishing industry connections, this agent handles the mechanical work while you focus on building genuine relationships.

Ideal for:

• Sales professionals who need to maintain consistent outreach volume
• Recruiters managing large candidate pools
• Business development teams scaling their networking efforts
• Marketing professionals conducting partnership outreach
• Anyone who values their time and wants personalized, not generic, connection requests

You should be comfortable with basic spreadsheet management and have access to LinkedIn profile URLs for your prospects. No coding knowledge required—just follow the setup steps below.

Required Integrations

Google Sheets

Why it's needed: Google Sheets serves as your central command center, storing prospect information, tracking connection status, and logging all generated messages. The workflow reads pending prospects, updates their status throughout the process, and records results.

Setup steps:

1. Enable Google Sheets API access in your Google Cloud Console

   • Visit Google Cloud Console
   • Create a new project or select an existing one
   • Navigate to "APIs & Services" > "Library"
   • Search for "Google Sheets API" and click "Enable"

2. Create service account credentials

   • Go to "APIs & Services" > "Credentials"
   • Click "Create Credentials" > "Service Account"
   • Name your service account (e.g., "TaskAGI LinkedIn Automation")
   • Grant it "Editor" role
   • Click "Done"

3. Generate and download JSON key

   • Click on your newly created service account
   • Navigate to the "Keys" tab
   • Click "Add Key" > "Create New Key"
   • Select "JSON" format and click "Create"
   • Save this file securely—you'll need it for TaskAGI

4. Configure in TaskAGI

   • Navigate to TaskAGI Integrations page
   • Click "Add Integration" > "Google Sheets"
   • Upload your JSON key file or paste its contents
   • Click "Authenticate" to verify connection
   • Name your integration (e.g., "LinkedIn Outreach Sheets")

5. Prepare your spreadsheet structure

   • Create a new Google Sheet with these columns:
     • Column A: Name (prospect's full name)
     • Column B: LinkedIn Profile ID (the profile identifier)
     • Column C: Status (values: "Pending", "In Progress", "Complete")
     • Column D: Generated Message (will be auto-populated)
     • Column E: Date Processed (optional, for tracking)
   • Share the sheet with your service account email (found in the JSON file as client_email)

Anthropic (Claude AI)

Why it's needed: Anthropic's Claude AI powers the personalization engine, analyzing LinkedIn profile data and crafting compelling, contextually relevant connection messages that feel genuinely human and significantly increase acceptance rates.

Setup steps:

1. Create an Anthropic account

   • Visit Anthropic Console
   • Sign up for an account or log in
   • Complete email verification

2. Obtain your API key

   • Navigate to "API Keys" in the console
   • Click "Create Key"
   • Name your key (e.g., "TaskAGI LinkedIn Outreach")
   • Copy the key immediately—it won't be shown again
   • Store it securely (consider using a password manager)

3. Configure in TaskAGI

   • Go to TaskAGI Integrations
   • Select "Add Integration" > "Anthropic"
   • Paste your API key in the authentication field
   • Click "Test Connection" to verify
   • Save your integration

4. Understand pricing

   • Claude Sonnet 4 costs approximately $3 per million input tokens
   • Each message generation uses ~500-1000 tokens
   • Budget approximately $0.003-0.005 per connection request
   • Set up billing alerts in Anthropic console to monitor usage

Configuration Steps

Step 1: Configure the Schedule Trigger

The Schedule Trigger node initiates your automation at regular intervals.

• Set interval to 1 minute for continuous processing
• Best practice: Start with a longer interval (5-10 minutes) during testing
• Once validated, adjust to your desired frequency based on daily outreach goals
• Consider LinkedIn's connection request limits (typically 100-200 per week for new accounts)

Step 2: Set Up Random Delay

The Random Delay node adds human-like variability to avoid detection as automated activity.

• Configure minimum delay: 1 minute
• Configure maximum delay: 5 minutes
• This randomization makes your activity appear more organic
• Important: Helps prevent LinkedIn rate limiting or account restrictions

Step 3: Configure Get Pending Prospect

The Get Pending Prospect node searches your Google Sheet for the next person to contact.

• Sheet URL: Paste your Google Sheet URL (the full URL from your browser)
• Search column: C (the Status column)
• Search value: Pending
• Return first match: Enable this option
• Include row data: Enable to capture all prospect information
• This node will return the first row where Status = "Pending"

Step 4: Set Up Mark as In Progress

The Mark as In Progress node updates the status to prevent duplicate processing.

• Sheet URL: Same as Step 3
• Row number: [[nodes.4114.matches.0.row_number]] (references the found row)
• Column: C
• New value: In Progress
• This immediately marks the prospect as being processed

Step 5: Configure Fetch LinkedIn Profile

The Fetch LinkedIn Profile node retrieves detailed profile information.

• HTTP Method: POST or GET (depending on your LinkedIn API access)
• URL: Your LinkedIn API endpoint or proxy service
• Body:

  {
    "profileId": "[[nodes.4114.matches.0.row_data.1]]"
  }
  

• Headers: Include authentication tokens for your LinkedIn API access
• Note: You'll need LinkedIn API access or a third-party service like PhantomBuster or Apify

Step 6: Configure Generate Personalized Message

The Generate Personalized Message node creates your connection request text.

• Model: claude-sonnet-4-5-20250929 (already configured)
• Prompt template:

  Generate a personalized LinkedIn connection message based on this profile data:
  
  Name: [[nodes.4114.matches.0.row_data.0]]
  Profile Info: [[nodes.4116.response]]
  
  Requirements:
  - Maximum 300 characters (LinkedIn limit)
  - Professional and friendly tone
  - Reference something specific from their profile
  - Clear value proposition for connecting
  - No generic templates
  - End with a question or call-to-action
  
  Return only the message text, no additional formatting.
  

• Max tokens: 200 (sufficient for connection messages)
• Temperature: 0.7 (balances creativity with consistency)

Step 7: Configure Send Connection Request

The Send Connection Request node submits the connection request to LinkedIn.

• HTTP Method: POST
• URL: Your LinkedIn connection request endpoint
• Body:

  {
    "profileId": "[[nodes.4114.matches.0.row_data.1]]",
    "message": "[[nodes.4117.response.content.0.text]]"
  }
  

• Headers: Include necessary authentication
• Error handling: Enable retry logic with 3 attempts

Step 8: Configure Mark as Complete

The Mark as Complete node updates the final status.

• Sheet URL: Same as previous Google Sheets nodes
• Row number: [[nodes.4114.matches.0.row_number]]
• Column: C
• New value: Complete

Step 9: Configure Update Message Column

The Update Message Column node logs the generated message for your records.

• Sheet URL: Same as previous Google Sheets nodes
• Row number: [[nodes.4114.matches.0.row_number]]
• Column: D
• New value: [[nodes.4117.response.content.0.text]]
• This creates an audit trail of all sent messages

Testing Your Agent

Initial Test Run

1. Prepare test data

   • Add 2-3 test prospects to your Google Sheet
   • Use LinkedIn profiles you control or test accounts
   • Set their Status to "Pending"

2. Execute manual test

   • In TaskAGI, click "Test Workflow"
   • Watch each node execute in sequence
   • Verify data passes correctly between nodes

3. Verify at each checkpoint:

   • ✅ Get Pending Prospect: Confirms it finds the first "Pending" row
   • ✅ Mark as In Progress: Check your sheet updates to "In Progress"
   • ✅ Fetch LinkedIn Profile: Review the returned profile data
   • ✅ Generate Personalized Message: Read the AI-generated message—is it personalized and relevant?
   • ✅ Send Connection Request: Verify the request appears in LinkedIn
   • ✅ Mark as Complete: Confirm status changes to "Complete"
   • ✅ Update Message Column: Check that the message is logged in column D

4. Expected results:

   • Total execution time: 15-30 seconds per prospect
   • All status updates reflected in Google Sheets
   • Connection request visible in LinkedIn "Sent" invitations
   • Generated message stored in your spreadsheet

Production Testing

1. Start with the schedule trigger disabled
2. Run 5-10 manual executions to validate consistency
3. Enable the schedule trigger with a 10-minute interval
4. Monitor for 1-2 hours to ensure smooth operation
5. Gradually reduce interval to your target frequency

Troubleshooting

"No pending prospects found"

Cause: The search isn't finding rows with "Pending" status.

Solutions:

• Verify column C contains exactly "Pending" (case-sensitive, no extra spaces)
• Check that your Google Sheet is shared with the service account email
• Ensure the sheet URL is correct and accessible
• Confirm the search column is set to "C"

"Failed to update Google Sheet"

Cause: Permission issues or incorrect row references.

Solutions:

• Re-share your Google Sheet with the service account (Editor permissions)
• Verify the row number reference: [[nodes.4114.matches.0.row_number]]
• Check that your service account JSON key is still valid
• Test the Google Sheets integration independently in TaskAGI

"LinkedIn API error" or "Connection request failed"

Cause: LinkedIn rate limiting, authentication issues, or API access problems.

Solutions:

• Verify your LinkedIn API credentials are current
• Check if you've hit LinkedIn's weekly connection limit (100-200 for newer accounts)
• Increase the random delay range (e.g., 5-15 minutes)
• Ensure your LinkedIn account is in good standing
• Consider using a LinkedIn automation proxy service

"AI message generation failed"

Cause: Anthropic API issues, insufficient credits, or prompt problems.

Solutions:

• Verify your Anthropic API key is valid and has available credits
• Check your Anthropic account billing status
• Review the prompt template for syntax errors
• Ensure profile data is being passed correctly from the previous node
• Test the Anthropic integration separately with a simple prompt

"Workflow processes same prospect multiple times"

Cause: Status update timing issues or failed status updates.

Solutions:

• Add a 2-3 second delay after "Mark as In Progress" node
• Verify the status update is completing before profile fetch begins
• Check Google Sheets API rate limits aren't being exceeded
• Ensure row number references are correct

Next Steps

After Successful Setup

1. Populate your prospect list

   • Import your target prospects into the Google Sheet
   • Ensure all have valid LinkedIn Profile IDs
   • Set initial Status to "Pending"
   • Start with 20-30 prospects for the first week

2. Monitor daily performance

   • Check your Google Sheet for completion rates
   • Review generated messages for quality and relevance
   • Track LinkedIn connection acceptance rates
   • Adjust your prospect criteria based on results

3. Establish a maintenance routine

   • Weekly: Add new prospects and review completed connections
   • Monthly: Analyze message effectiveness and refine prompts
   • Quarterly: Review LinkedIn account health and connection limits

Optimization Suggestions

Improve message quality:

• Refine your Claude prompt with specific industry language
• Add company research data to the profile fetch step
• Include recent posts or activity in the personalization context
• A/B test different message styles by creating workflow variants

Scale responsibly:

• Gradually increase daily volume (start with 10-15 per day)
• Monitor LinkedIn account health indicators
• Diversify prospect sources to avoid pattern detection
• Implement time-of-day scheduling for peak engagement

Enhance tracking:

• Add a "Date Sent" column with timestamp: [[nodes.4119.timestamp]]
• Create a "Response Received" column for manual tracking
• Build a dashboard to visualize connection acceptance rates
• Set up email notifications for daily completion summaries

Advanced Usage Tips

Segment your outreach:

• Create multiple sheets for different prospect categories
• Duplicate the workflow for each segment with tailored prompts
• Use different Claude temperature settings for various industries
• Implement priority queues with multiple "Status" values

Add follow-up sequences:

• Create a second workflow that processes "Connected" prospects
• Send follow-up messages 3-5 days after connection acceptance
• Build a nurture sequence with valuable content sharing
• Track engagement and move hot prospects to your CRM

Integrate with your sales stack:

• Connect to your CRM to sync new connections automatically
• Trigger Slack notifications when high-value prospects connect
• Export weekly reports to Google Data Studio
• Build webhooks to notify your sales team of warm leads

Implement safety measures:

• Set daily maximum limits using a counter variable
• Create a "pause" mechanism for weekends or holidays
• Add profile quality filters to skip incomplete LinkedIn profiles
• Implement a "cooldown" period between prospects from the same company

You're now ready to transform your LinkedIn outreach! Start small, monitor closely, and scale gradually. Remember: the goal is quality connections, not just quantity. Let the automation handle the mechanics while you focus on building meaningful professional relationships.