⚡ One-click install
Checked install page (app stores, official sites and template pages): tap to get it straight away.
🧰 Before you start
- Claude desktop app with a paid plan (download button on this page)
- Free Apify and Apollo accounts (Apify collects public post comments; Apollo looks up job titles and companies)
- A one-paragraph description of your ideal customer
🪜 Step by step
- 1☐
Install Claude desktop and open Cowork
Use the download button on this page, sign in, and look for the Cowork tab (where Claude does longer jobs on your computer).
- 2☐
Describe your ideal customer
Write who you want: job titles, company size, location and what problem they have. For property, e.g. 'Malaysian professionals interested in overseas property investment'.
- 3☐
Find a busy post in your field
Pick a LinkedIn post on your topic with lots of comments, ideally from the last week. Copy its link.
- 4☐
Collect the commenters
Ask Claude to use Apify (connect it when asked, using your Apify API key) to pull the names and comments. If you'd rather not connect it, export the comments from Apify yourself and upload the file.
- 5☐
Look up and score them
Ask Claude to add each person's role and company (from Apollo or their public profile) and score them 1-5 against your ideal customer, with a one-line reason.
- 6☐
Get drafts for the top scorers
For people scoring 4-5, ask for a short opening message that mentions their actual comment. Read and edit each one; you send them yourself.
- 7☐
Save it as a skill
Ask Claude to save the whole process as a skill (a reusable recipe) called 'Comment lead pipeline' so next time you just paste a new link.
💬 Copy-paste prompts
Tap Copy prompt, paste it into your AI (Grok, ChatGPT, Claude or Gemini), then replace the [yellow bits] with your own details.
Here is a LinkedIn post: [post link]. My ideal customer: [job titles, company size, location, problem they have]. 1. Collect everyone who commented, with their comment text. 2. For each person, add their role and company from public information. If you can't find it, write 'unknown'. Don't guess. 3. Score each 1-5 for fit with my ideal customer and give a one-line reason. 4. Save a spreadsheet sorted by score. 5. For everyone scoring 4 or 5, draft a 3-sentence opening message (British English) that refers to what they actually said and offers [my offer]. Don't send anything.
Save the process we just did as a skill called 'Comment lead pipeline'. It should ask me for a post link and my ideal-customer description, then run the same steps and produce the same spreadsheet and drafts.
⚠️ Tips & common mistakes
- Only use public information and follow LinkedIn's rules; heavy automated collecting can get accounts restricted.
- Under Malaysia's PDPA, don't keep personal data longer than you need, and let people opt out.
- Message a handful of people personally rather than blasting everyone.