Email Pattern Finder — staff names into addresses

Companies publish their team and hide their addresses. One accountancy firm in testing listed 41 named staff and zero emails. Those names are worthless to a salesperson until you can address a message to them.

This crawls the site, collects any published addresses on the company's own domain, works out which pattern the company uses, and applies it to everyone else.

How it decides

It tests every observed name → address pair against 11 common patterns (first.last, flast, f.last, firstlast, last.first…) and counts agreement. Confidence is reported honestly:

evidence	confidence
3+ matching published addresses	high
2	medium
1	low — one match can be coincidence
0 personal addresses published	not inferable — you get ranked guesses, labelled as guesses

Only addresses on the company's own domain are used. An agency or partner address would infer the wrong pattern entirely.

Works outside English

Job titles, contact-page names and personal names are matched in English, Spanish, German, French, Italian, Portuguese and Dutch. An English-only version found zero people on every Spanish site tested — which reads as "this firm publishes no staff" when the truth is we were not looking at the right page.

It also handles Spanish and Portuguese double surnames. "José García López" is addressed jose.garcia@, using the paternal surname — not jose.lopez@. Taking the last name made every such firm un-inferable; both conventions are now tested and the winning one is reported (first.last|paternal).

What it does not do

It does not verify deliverability. SMTP probing is intrusive, unreliable and gets your infrastructure blacklisted. Every address is marked published, inferred or guess so you know exactly what you have. Anyone claiming verified addresses from a crawl alone is overstating it.

Output

{
  "domain": "example.co.uk",
  "pattern": "first.last", "patternConfidence": "high",
  "patternEvidence": [{ "name": "Sarah Jones", "email": "sarah.jones@example.co.uk" }],
  "peopleCount": 41, "addressableCount": 41,
  "contacts": [
    { "name": "David Hardie", "title": "Accountant",
      "email": "david.hardie@example.co.uk", "source": "inferred", "confidence": "high" }
  ]
}

The pipeline

Local Business Finder → Lead Contact Finder → Email Pattern Finder
     companies              named people           addresses

Each stage is useful alone; together they turn a town and a trade into a contact list.