Find companies no database carries
Source from certification registers, trade associations, licence lists, PDFs and government portals, then verify every row against your niche.
3 steps not yet verified live
Tool names and credit costs in this recipe are validated. Some steps work around a platform gap that has not been decisively probed, so treat those as a workaround rather than a documented capability. Check the result before you act on it.
Where the list comes from
Fragmented public sources: registries, associations, licence lists, portals and documents. The ICP is converted into search queries first.
The steps
- 1
Orient
Confirm the inputs, compute the expected cost, and tell the user what the run will do before anything spends a credit.
- 2
Convert the ICP into source-search queries
Before scraping anything, work out where these companies are listed. This is a research step, not a data step, and it decides the quality of everything after.
- 3
Scrape each source
Scrape each source into its own project, then merge. A merged scrape is impossible to debug when one source turns out to be garbage.
- 4
Clean extracted rows, remove non-company entries
WorkaroundDirectory scrapes always contain headers, category labels, pagination text and adverts masquerading as rows. no "is this row actually a company?" validator. Use AI - System Prompt (0.5c).
- 5
Normalize and deduplicate
Clean Company Names (0.1c) then Clean Location (0.1c).
- 6
Resolve domains
Official Website Finder (1c). Then Normalize Website URL (0.1c) and Check Domain Status (FREE) to drop parked and dead domains. Directory-sourced companies have a materially worse domain hit rate than database-sourced ones.
- 7
Enrich firmographics
Industry Classifier (0.2c), Company Headcount (0.2c), Clean Location (0.1c), Business Model Finder (0.2c). Description: Custom Research Tool (1c) as in SOP 01 Step 5. Gate: enough populated data for Step 7 to judge against.
- 8
AI verification against the niche ICP
Scraping from a "certified companies" directory does not prove certification. Verify explicitly, with evidence. Custom Research Tool (1c).
- 9
Run the shared outbound tail
See _shared-outbound-tail.md. Personas: whoever owns the niche attribute, for a certification ICP that is Compliance / Quality / Operations, not Sales. Signal column: the Step 7 evidence string.
Where the run should stop
- Non company entries are removed before domains are resolved.
- AI verification against the niche runs before any contact enrichment.
What you should end up with
A deduplicated, domain resolved list where every row has been verified as genuinely in your niche, with the source it came from attached.