4-Platform Lead Scraper — Prompt Pack
Copy-paste AI prompts to extract and qualify B2B leads from LinkedIn, Google Maps, company sites, and niche directories — without buying another SaaS scraper.
What this pack does
You do not need a $200/mo scraper to build a first-pass lead list. These four workflows use AI plus public pages you can open manually (or via your own compliant automation). Output is always a structured table you can paste into Sheets, Airtable, or your CRM.
Compliance note: Scrape only what you are allowed to access. Respect robots.txt, platform Terms of Service, and GDPR/CAN-SPAM when you email anyone. This guide is for research and outbound prep — not spam.
The four platforms
| Platform | Best for | Typical fields |
|---|---|---|
| Decision-makers, titles, company size | Name, title, company, geo | |
| Google Maps | Local service businesses | Business, phone, site, reviews |
| Company websites | Emails, team pages, tech stack | Contact, about, pricing signals |
| Niche directories | Vertical lists (Clutch, G2, associations) | Vendor name, category, region |
Workflow 0 — ICP brief (run once)
Before scraping, lock your ideal customer profile:
Define my ICP:
- Industry: [e.g. commercial HVAC]
- Geo: [city / state / remote]
- Company size: [employees or revenue band]
- Buyer title: [owner, ops manager, etc.]
- Disqualifiers: [franchise HQ only, no website, etc.]
Output: 5-bullet ICP + 10 keyword phrases I should search.
Platform 1 — LinkedIn search export
- Run a LinkedIn search (People or Companies) with your ICP filters.
- Copy the visible results page text (or export via Sales Navigator if you have it).
- Paste into AI:
Extract leads from this LinkedIn results paste.
Columns: full_name | title | company | location | profile_url | fit_score_1-5 | fit_reason
Rules:
- Skip duplicates
- fit_score 1 = wrong industry, 5 = perfect ICP match
- If field missing, write UNKNOWN
Max rows: 50
[PASTE]
Platform 2 — Google Maps local list
- Search Maps for
[service] in [city]. - Scroll and copy business names, ratings, and snippets (or use Google's export if available).
- Prompt:
Turn this Google Maps paste into a lead table.
Columns: business_name | category | address | phone | website | rating | review_count | outreach_angle
outreach_angle = one sentence why they might need [your offer]
Flag businesses with no website as "digital gap"
[PASTE]
Platform 3 — Company site contact mining
For each target domain, open /about, /team, /contact. Copy visible text.
From this company website paste, extract:
- company_name | primary_contact_name | role | email | phone | linkedin_hint
- tech_signals (CMS, chat widget, booking tool if visible)
- personalization_hook (one specific fact for cold email)
Do not invent emails. If not on page, mark email: NOT_FOUND
[PASTE]
Platform 4 — Directory sweep
Pick one directory your buyers use (Clutch, G2, industry association member list). Copy listing blocks.
Parse directory listings into:
company | category | location | website | size_hint | directory_url | icp_match_y/n | notes
Sort icp_match y first. Dedupe by domain.
[PASTE]
Merge and dedupe prompt
After all four passes:
Merge these four CSV tables. Dedupe by company domain (normalize www).
Add column: source_platform
Add column: priority (high/medium/low) using ICP: [paste ICP]
Output final CSV only, max 100 rows, high priority first.
[PASTE TABLES]
Qualification checklist
- ICP written before any search
- Every row has a source (URL or platform)
- No invented emails or phone numbers
- fit_score or priority applied
- Personalization hook ready for top 20 rows
Outreach prep (optional next step)
For these 10 high-priority leads, write a 3-line cold email each:
Line 1: specific hook from their site/reviews
Line 2: one outcome we deliver
Line 3: soft CTA (15-min call / reply yes)
Tone: direct, no hype. Max 90 words each.
[PASTE TOP 10 ROWS]
More free operator playbooks live in the MyGearHut library. Join The Gear Drop if you want the next prompt pack dropped in your inbox.