Step-by-step guide: Generate 5,000+ personalized emails based on LinkedIn data
Attention: This guideline teaches you to run ChatGPT4 models to generate thousands of personalized emails. Personize is the first and only no-code studio for GTM teams that gives you the power to run the most advanced Gen AI models like ChatGPT at scale and with a few clicks.
Are you ready? Let's start. 🔥
First, watch this video and then apply the following step-by-step guidelines to your own CRM data. 👇
To generate your first batch of personalized emails, you must follow three steps as follows:
Step 1: Upload Your Contact List
Upload a CSV file containing your contacts. Ensure each entry includes the first name, title, company name, LinkedIn Person URL, and LinkedIn Company URL.
Add additional columns as memory.
Sync data from Google Sheets
Sync data from HubSpot
Step 2: Build Your Skills
What is a Skill in Personize Studio?
Skill is designed to receive a specific set of inputs as dynamic variables from every row of a table and then generate an output, typically generated by AI.
You can design and add multiple skills for your business's diverse tasks, such as one for composing the initial outbound email, another for crafting a follow-up message, and yet another for drafting LinkedIn messages. The number of skills you can create is unlimited, allowing for greater flexibility and customization.
Visit the Skills tab and click '+' to create a new skill.
Set up your skill: you can use the following System Message and Prompt.
System Message (Instruction):
The system message provides a framework for LLMs to process your skill prompts. You can assign a persona to your skill and also add instructions on how to process the data of each row in the prompt.
Act as an experienced salesperson who is known for crafting compelling emails to c-level executive buyers.
You will get requests from the Sender to write emails to a given lead. You always meticulously follow these steps:
Step 1) read the information provided about the lead's company or her/his personal LinkedIn data.
Step 2) patiently analyze the Sender's pitch and then use your creativity personalize the pitch based on what you learned in step 1 about the given lead and their company.
Step 3) write the whole email and personalize the first paragraph based on the personalized pitch you developed in previous step.
Step 4) This is your final step and you thoroughly quality check your email draft to ensure highest click rate and positive reply rate.
You always listen to the instruction provided by the sender to you in their requests.
Dynamic Prompt:
Now, it is time to add your prompt. The dynamic prompt contains variables for which they will be replaced by real values for each row. Please use your own pitch and Sender name below:
Write an email to {{first name}}, the {{title}} at {{company name}}.
Sender: Hamed, CEO at Personize.ai
Personalize this email based {{first name}}'s LinkedIn data:
"""
{{lead's LinkedIn data}}
"""
Pitch: Have you explored Generative AI like ChatGPT for creating engaging emails?
CTA: Book a Meeting.
Tone: Friendly, casual.
Limit: Under 100 words.
Okay! I want to invite you to build your second skill yourself.
Second Skill: Write a follow up email that will be sent if and only if the lead does not respond to the first email within a time frame. Personalize this follow up email based on the lead's company data.
Step 3: Go to the Runs Tab and Create your First Run
◾ Click '+' to start a new run. Select your CSV file.
We recommend starting with a small range (e.g., 5 contacts) to check the quality of the generated emails before continuing.
◾ Add the LinkedIn Enrichment Tool and link it to the Person and Company LinkedIn URLs in your database. You need two steps. In each step, you use this tool for one of the URL columns.
Personize has its own tools to enable you to read the web and also LinkedIn data for a given contact or company.
◾ Choose the "Outbound Email Writer" skill. Map your CSV columns to the corresponding fields in the skill.
Select a Provider: Select OpenAI and your preferred model (e.g., GPT-4, ChatGPT 3.5 Turbo). Keep default settings for temperature and stopper.
◾ Choose the "Follow Up Email Writer" skill. Map your CSV columns to the corresponding fields in the skill.
Select a Provider: Select OpenAI and your preferred model (e.g., GPT-4, ChatGPT 3.5 Turbo). Keep default settings for temperature and stopper.
◾ Choose to update the existing table. Run.
It only takes a few seconds. When you run is in the completed tab, it is ready for your review.
Note: When you are building your first run, you need to connect to your OpenAI account. Get the OpenAI API key here:
Congrats! You just learned how to use Personize to generate personalized emails.
You may ask: "What happens during your Run?"
For each row, Personize will scrape the website content from the provided URLs and, then insert the the scrapped content and other input variables into the dynamic prompt inside the "Outbound Email Writer" and then "Follow up Email Writer" skills. Then it generates personalized emails using the ChatGPT engine.
Step 4: Review, Optimize, and Re-Run
Once satisfied with a small batch, scale up by increasing the range to 1000s of contacts or more for larger runs.
Step 5: Download CSV or JSON file or Sync via HubSpot/Google Sheets integrations
Go back to the data tab, choose your file/table, and then you can download the CSV or JSON file. Alternatively, you can go to the Integrations tab and sync the results back to HubSpot or to your Google Sheets.