LinkedIn Lead Personalization: Google Drive + Apify + AI Cold Outreach Opener

Upload a Google Drive spreadsheet → parse rows and scrape LinkedIn activity → generate a one-sentence personalized opener with an LLM → write back to Google Sheets and send a Telegram summary

Who

Sales/SDR teams, lead-gen agencies, growth marketers

You get

Auto-generate a single personalized cold-email opener per lead based on their latest LinkedIn post or profile, and write it back to the original sheet

You need

n8n · google-drive · google-sheets · apify · anthropic · openai · telegram

Steps

1. Connect credentials and configure trigger

   In n8n, connect: Google Drive/Sheets (OAuth2), Apify (token), Anthropic and/or OpenAI (API key), Telegram (bot token). Open the Google Drive Trigger node and select the folder to watch (triggers when a new spreadsheet is uploaded).

2. Prepare spreadsheet and map columns

   Ensure the uploaded sheet contains these columns: First Name, Last Name, Company Name for Emails, Person Linkedin Url, Email, Website. The Code1 node maps them to firstName, lastName, company, linkedinUrl, email, website.

3. Replace placeholders

   In the HTTP Request nodes (Scrape LinkedIn Profile Post, Colect LinkedIn Data), replace the Authorization header with Bearer your actual Apify token. In the Telegram node (Lead Enrichment is ready), replace chatId with your actual Telegram Bot Chat ID.

4. Scrape LinkedIn and check recency (OpenAI)

   For each contact, the workflow calls Apify LinkedIn Profile Posts to fetch the latest post, while the Date & Time node records the current date. The OpenAI node (gpt-4.1-mini) evaluates whether the post is from the current year, returning only true or false.

   Execution prompt · Source text。OpenAI node system message field

   Copy prompt

   You are a LinkedIn Post Analyzer that determines if a profile has posts from the current year. Your sole purpose is to examine LinkedIn post data and determine if there are any posts from 2025.
   TASK:
   Analyze the provided LinkedIn profile post data and determine:
   
   If the profile has ANY posts
   If any posts are from the current year (2025)
   
   OUTPUT FORMAT:
   You must respond with EXACTLY ONE of these values:
   
   "true" - If there is at least one post from 2025
   "false" - If there are no posts from 2025
   
   Do not include any explanations, reasoning, or additional text - ONLY return "true" or "false" based on your analysis.

   Execution prompt · Source text。OpenAI node user message field

   Copy prompt

   =Examine the following LinkedIn profile data and determine if there are any posts from the current year - 2025.
   LinkedIn Profile Data:
   
   Post: {{ $('Scrape LinkedIn Profile Post').item.json.text }}
   Date: {{ $('Scrape LinkedIn Profile Post').item.json.posted_at.date }}
   Current Date: {{ $json.currentDate }}
   
   If the data contains at least one post with a timestamp or date indicating it was posted in 2025, respond with only the word "true".
   If the profile has no posts OR all posts are from years prior to 2025, respond with only the word "false".
   Your entire response must be ONLY "true" or "false" with no additional text.

5. Branch A: Generate opener from recent post (Anthropic)

   If OpenAI returns true, the AI Agent1 node calls the Anthropic Chat Model (claude-sonnet-4-20250514) to craft a single 25-30 word natural reference sentence based on the recent LinkedIn post.

   Execution prompt · Source text。AI Agent1 node user prompt field (text)

   Copy prompt

   =I need to create a personalized outreach message for a prospect. Please use this recent LinkedIn post from their profile and generate a single personalized sentence I can use in my outreach email.
   
   LinkedIn Post:
   {{ $('Scrape LinkedIn Profile Post').item.json.text }}
   
   Context: I own an AI automation agency focused on building Lead Generation Outreach and Lead Enrichment Systems. I'm reaching out to this person because their work appears relevant to our services.
   
   Generate a single personalized sentence that references their LinkedIn activity in a natural, non-intrusive way that I can include in my outreach email.

   Execution prompt · Source text。AI Agent1 node system message field (options.systemMessage)

   Copy prompt

   You are a LinkedIn Post Analyzer specialized in creating personalized outreach sentences. Your purpose is to examine the most recent LinkedIn posts from a prospect's profile and generate a single, authentic engagement sentence that can be used in cold email outreach.
   
   TASK:
   1. Use the most recent LinkedIn post provided in the user prompt.
   2. From the selected post, craft ONE personalized sentence (25-30 words maximum) that:
      - References specific content from the post without appearing stalkerish
      - Creates a genuine connection based on the post's content
      - Uses conversational language that sounds natural in an email
      - Avoids generic compliments or obvious flattery
      - Maintains a professional tone while being personable
   
   IMPORTANT: Your output must ONLY contain the single personalized sentence with no labels or prefixes. Do not include "Personalized Sentence:" or any other text before your response. Do not include greetings, introductions, or closings. Focus solely on creating a natural reference to their LinkedIn post that demonstrates genuine interest in their professional perspective.
   
   The output will be inserted directly into a CRM system with no human review, so it must be ready to use without any additional formatting or labels.

6. Branch B: Generate opener from profile data (Anthropic)

   If OpenAI returns false, the workflow first calls the Apify LinkedIn Profile Scraper to retrieve headline, about, and experiences, then the AI Agent_Personalization_no LN post node calls Anthropic to generate a 15-25 word cold-email opener.

   Execution prompt · Source text。AI Agent_Personalization_no LN post node user prompt field (text)

   Copy prompt

   =Create ONE personalized, emotional opening sentence for a cold email to {{ $json.firstName }} {{ $json.lastName }} that will compel them to read further. Focus on our Lead Generation Sales System that saves time.
   Output only the sentence, nothing else.

   Execution prompt · Source text。AI Agent_Personalization_no LN post node system message field (options.systemMessage)

   Copy prompt

   =You are a sales development expert crafting personalized cold email opening lines without website data.
   
   Your task is to create ONE compelling opening sentence for a cold email that will grab attention immediately. focus on:
   
   1. Using the name of the lead: {{ $json.firstName }}
   2. Using their headline: {{ $json.headline }}
   3. Using their LInkedIN about section: {{ $json.about }}
   4. Current company: {{ $json.experiences[0].subtitle }}
   5. Addressing a universal pain point related to lead generation that resonates with their likely business challenges
   6. Creating curiosity about our Lead Generation Sales System
   7. Mentioning a specific benefit (saving time, reducing costs, increasing conversion rates)
   8. Making it conversational and emotionally engaging (15-25 words maximum)
   
   OUR PRODUCT: We offer a Lead Generation Sales System that automates prospect qualification, outreach sequences and lead enrichment after booking. Key benefits include:
   - 60% reduction in time spent on manual prospecting
   - 40% lower customer acquisition costs
   - 3x higher quality leads through AI-powered targeting
   - Elimination of repetitive sales tasks
   
   TONE GUIDELINES:
   - Confident but not aggressive
   - Curious rather than presumptive
   - Focused on their potential gains, not their current problems
   - Conversational, as if from one professional to another
   
   EXAMPLES OF EXCELLENT OPENERS:
   - "What if your sales team at [Company] could focus exclusively on closing while an automated system handles all prospecting, [First Name]?"
   - "The companies outperforming their competitors this quarter have one thing in common, [First Name]: they've eliminated manual lead generation entirely."
   - "Imagine cutting your customer acquisition costs by 40% while simultaneously improving lead quality, [First Name] - that's what our system recently did for a company similar to [Company]."
   - "While researching innovative companies in the [derived industry] space, [Company]'s approach stood out as perfectly positioned to benefit from automated lead generation."
   
   Remember: ONE sentence only. Make it compelling enough that they WANT to read the next line of the email.

7. Write back to Google Sheets and loop

   The Personalization node assigns the AI output to the personalization field, along with sheetId, FileName, and SheetName. The Google Sheets Update with Personalization node matches on the email column and writes personalization back to the original sheet. The Loop Over Items1 node processes rows one by one; after all rows finish, execution proceeds to End of Loops.

8. Send Telegram completion notification

   The End of Loops node (JavaScript Code) aggregates results from all rows,拼接 the sheet name and link, and sends a Markdown-formatted message to Telegram.

This template is an n8n workflow setup guide. Follow this order: 1) Import the workflow JSON into n8n; 2) Connect Google Drive/Sheets, Apify, Anthropic/OpenAI, and Telegram credentials; 3) Select the Drive folder to watch; 4) Confirm sheet column names match the template; 5) Replace placeholders in HTTP and Telegram nodes; 6) Upload a test spreadsheet and execute once to validate.