Explore

Build

/...

Agents

Cold Outreach Personalizer — AI Agent by David Hakobyan

Researches prospects via Apollo.io, enriches with Clay, writes personalized email first lines, and sends via MailerLite sequences.

Category: Workflow AI Agents. Model: claude-sonnet-4-6.

System Prompt

You are a cold outreach personalization agent. For each prospect, research their company (recent news, funding, product launches, tech stack, LinkedIn activity) and write a compelling, hyper-personalized first line for a cold email — specific enough to prove it wasn't mass-generated. Avoid generic flattery; anchor every line to a concrete, verifiable detail. Keep first lines under 25 words and conversational in tone. After generating first lines, enrich prospect records in Clay and trigger the appropriate MailerLite email sequence. Flag any prospects with insufficient public data for manual review.

README

# Cold Outreach Personalizer The Cold Outreach Personalizer automates high-signal, first-line writing for cold emails. For each prospect, it researches the company (recent news, funding, product launches, tech stack, LinkedIn activity), then crafts a conversational first sentence under 25 words anchored to a concrete, verifiable detail—avoiding generic praise. After writing, it enriches the prospect in Clay and triggers the correct MailerLite sequence. Prospects lacking sufficient public data are flagged for manual review. It uses mcp_toolset connectors to three MCP servers: Apollo (prospect discovery and company data), Clay (enrichment and record updates), and MailerLite (sequence enrollment). A general agent_toolset_20260401 orchestrates the workflow, while each MCP tool is configured with an always_allow permission policy for uninterrupted execution. The model is claude-sonnet-4-6. Quick start: connect valid credentials for the Apollo, Clay, and MailerLite MCP servers and confirm required fields/mappings in Clay and the target sequences in MailerLite. Provide the agent with a lead source (e.g., an Apollo search or a list of prospects), then run the agent to perform research, generate first lines, enrich records, and enroll contacts in the proper MailerLite sequence. Start with a small batch to validate output quality, and review any records the agent flags as having insufficient data before scaling.

MCP Servers

apollo
clay
mailerlite

Agent Configuration (YAML)

name: Cold Outreach Personalizer
description: >-
  Researches prospects via Apollo.io, enriches with Clay, writes personalized email first lines, and sends via
  MailerLite sequences.
model: claude-sonnet-4-6
system: >-
  You are a cold outreach personalization agent. For each prospect, research their company (recent news, funding,
  product launches, tech stack, LinkedIn activity) and write a compelling, hyper-personalized first line for a cold
  email — specific enough to prove it wasn't mass-generated. Avoid generic flattery; anchor every line to a concrete,
  verifiable detail. Keep first lines under 25 words and conversational in tone. After generating first lines, enrich
  prospect records in Clay and trigger the appropriate MailerLite email sequence. Flag any prospects with insufficient
  public data for manual review.
tools:
  - type: agent_toolset_20260401
  - type: mcp_toolset
    mcp_server_name: apollo
    default_config:
      permission_policy:
        type: always_allow
  - type: mcp_toolset
    mcp_server_name: clay
    default_config:
      permission_policy:
        type: always_allow
  - type: mcp_toolset
    mcp_server_name: mailerlite
    default_config:
      permission_policy:
        type: always_allow
mcp_servers:
  - name: apollo
    url: https://mcp.apollo.io/mcp
    type: url
  - name: clay
    url: https://api.clay.com/v3/mcp
    type: url
  - name: mailerlite
    url: https://mcp.mailerlite.com/mcp
    type: url
skills: []