Check prospect fit and write cold email openers
Claude Code automates checking whether companies fit the target and creates personalized openers, saving manual prospect checking. Task: Suggested implementation: Export a broad company list from Apollo. Use Claude Code to build a script that scrapes each company's website and labels it 'yes' or 'no' based on your Ideal Customer Profile (ICP). Remove 'no' companies, re-import 'yes' companies into Apollo, and extract lead emails. Human role: Define ICP criteria, review AI labels for accuracy, and approve final lists.; tools: Claude Code. Human role: The original post does not say how a person stayed involved.; limitations: Prerequisites: Apollo account, Claude Code environment, and clear ICP definitions. Constraints: Web scraping may fail on dynamic sites or be blocked by robots.txt. Cost and time are unspecified. Human approval is required before sending emails to avoid compliance issues. Source reliability is low due to access errors..
Tools
Published on Agent Work Atlas:
Overview
Claude Code automates checking whether companies fit the target and creates personalized openers, saving manual prospect checking.
Workflow
Suggested implementation: Export a broad company list from Apollo. Use Claude Code to build a script that scrapes each company's website and labels it 'yes' or 'no' based on your Ideal Customer Profile (ICP). Remove 'no' companies, re-import 'yes' companies into Apollo, and extract lead emails. Human role: Define ICP criteria, review AI labels for accuracy, and approve final lists.
Reported result
Actual outcomes are unknown as the source is unavailable (HTTP 403). Proposed measurable acceptance check: Calculate the percentage of AI-labeled 'yes' prospects that are manually verified as valid fits. Target: >80% accuracy to justify automation.
Human role
The original post does not say how a person stayed involved.
Limitations
Prerequisites: Apollo account, Claude Code environment, and clear ICP definitions. Constraints: Web scraping may fail on dynamic sites or be blocked by robots.txt. Cost and time are unspecified. Human approval is required before sending emails to avoid compliance issues. Source reliability is low due to access errors.
Try it
I need to automate prospect qualification. First, export 50 companies from Apollo to a CSV file [replaceable input: Apollo export]. Use Claude Code to write a Python script that visits each company's website, checks for keywords matching my ICP [replaceable input: ICP keywords], and labels them 'yes' or 'no'. Run the script on this small batch. Output a CSV with columns: Company Name, Website, AI Label, Reason. I will manually review the 'yes' labels. If accuracy is below 80%, adjust the prompt. Do not send emails yet. Confirm the script works on the sample before scaling to 500 companies. Handle errors by logging failed URLs.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.