Business Enrichment - Company Profiles and Contacts

Turn a list of business names (or website URLs) into source-backed company records: official website, published emails and phone numbers, and a description of what the business does. Built for supplier lists and incomplete CRM records, including small businesses that have no website in your data.

Every fact in the output carries the URL it came from. Nothing is generated or guessed: if a fact is not published on a page that matches the business, the field stays empty.

What you get

• Identity match with a transparent confidence level and the reasons behind it (name in page title, city, country, matching phone).
• Official website, only when it is confirmed: supplied in your input, or a branded domain that also shows your phone number.
• Emails and phones published by the business, first-party sources first. Directory publishers' and article authors' mailboxes are filtered out.
• Business description: an excerpt of published activity or background text, never a generated summary.
• Optional email verification through LeadProof Bulk Email Verifier.
• One output row per input row, including businesses that could not be found, so your export lines up with your source list.
• Pay only for matches. Unresolved rows are free. A match is an identified business, not a guaranteed email (see Pricing).

Input

{
  "businesses": [
    {"id": "supplier-1", "name": "Eon Energy", "country": "Kenya", "phone": "+254202210028"},
    {"id": "supplier-2", "name": "Kavirondo Chemist", "country": "Kenya", "city": "Homa Bay"},
    {"id": "supplier-3", "website": "bidcoafrica.com", "country": "Kenya"}
  ],
  "verifyEmails": false
}

Each row needs a name or a website. Name searches also need a country, per row or through the top-level country default. Optional fields: id, city, phone, email, address, category. Keep phone numbers as strings so leading zeroes survive.

• City narrows the search and supports the identity check. If the first search returns no readable page, the Actor retries once without the city.
• Phone is the strongest identity signal: a page showing your phone number next to the business name is a high-confidence match.
• Email, address and category are kept as context in input. They are never presented as discovered facts.

Instead of businesses you can pass csvUrl: a public or signed UTF-8 CSV (up to 2 MB, 100 rows) with the same column names. Up to 100 businesses per run (maxBusinesses). Duplicate rows are processed separately so each keeps its own id.

Output

Flat columns for spreadsheets: inputId, businessName, status, matchConfidence, websiteUrl, businessDescription, emailAddresses, phoneNumbers, warnings. Use the Download CSV link on the run page for exactly these columns.

The JSON adds the evidence behind each value:

• website, description, emails, phones, addresses, socialProfiles, productsAndServices, businessTypes: each fact with sourceUrl, confidence and method.
• emails[].contactScope: city_named_mailbox when the mailbox is named after your row's city (for example a branch mailbox), otherwise company_or_listing.
• emails[].verificationStatus and isDeliverable when verification is on.
• sources (every page considered, with match reasons) and candidates (pages that did not pass the identity check, for manual review).
• input: your original row.

Statuses

Status	Meaning	Charged
matched	At least one readable page carries the business name plus the city, country or phone.	Yes, one business-match
needs_review	Candidate pages mention the name or phone, but none passed the identity check.	No
not_found	Search returned nothing that mentions the business.	No
error	The search dependency failed. This says nothing about the business.	No

Confidence and scope

matchConfidence is a transparent rule, not a calibrated probability:

• high: supplied website, or page title with the name plus your phone number.
• medium: a distinctive name in the page title plus your city or country. Review medium matches before outreach.

A company match does not guarantee the right branch or a named person. Large companies publish head-office and other branch contacts; the Actor ranks a mailbox named after your city first and adds the warning contacts_may_include_head_office_or_other_branches. A country-only match can confuse two businesses with the same name. Generic names (for example "General Supplies") are never matched on location alone.

Verification semantics

With verifyEmails: true, each found email is checked. isDeliverable is true only for an ok verdict and false only for invalid. catch_all, unknown and provider errors stay null: they do not confirm deliverability. With verification off, no verification runs and every email is not_requested.

Pricing

$0.01 per matched business (business-match). Rows with needs_review, not_found or error are free. The only other event is the standard Actor start ($0.00005).

A match is not a guaranteed email. business-match is charged when the business is identified on a public page. A matched row can have an email, a phone, a website and a description, or only some of them: many small businesses publish a phone number but no email. Check emailAddresses and warnings per row.

Dependency runs are billed in addition to the $0.01. This Actor calls other Actors on your account, and each bills you separately:

Dependency	When it runs	Billed by it
RAG Web Browser (Apify)	One search per business, a second only when the first returns no readable page	Platform usage: Google SERP proxy plus compute, typically under $0.01 per search
Website Email Finder (LeadProof)	Only for a confirmed website, when search pages did not fill the email quota	Its Store price per email found
Bulk Email Verifier (LeadProof)	Only with verifyEmails: true	Its Store price per email checked

Your maximum charge for the run applies to business-match events only. The Actor stops starting new searches once the remaining limit cannot cover every business in progress. Each dependency run has its own cap of $0.20. Start with a few rows to see real costs for your list.

Example total cost

Scaled from our pilot on a list of small Kenyan suppliers, verification off, per 100 input rows:

Item	Billed by	Approximate cost
About 40 matched businesses x $0.01	This Actor	$0.41
Searches, including retries	RAG Web Browser (platform usage)	$0.55
About 11 emails found on websites x $0.005	Website Email Finder	$0.06
Total		about $1.02

With verifyEmails: true, add about $0.0049 per email found (about $0.16 for this list). Search usage follows your Apify plan's platform prices, and your total depends on your match rate and how many emails businesses publish.

Coverage limits

• Only public pages returned by the search are read. There is no login and no dedicated Facebook, LinkedIn or Google Maps lookup; public pages from those sites count only when the search returns them readable. Blocked or JavaScript-only pages stay unresolved.
• Small local businesses often have no web presence at all. Expect many not_found rows on lists of sole traders.
• Search results vary between runs, so rerunning an unresolved row can find something new.
• Directories can be stale or mix several businesses. On directory pages, only company-branded emails or the email you supplied are accepted, which deliberately skips some Gmail or Yahoo addresses to avoid publisher contamination.
• Addresses and products/services are filled only from structured business data on the page, so they are often empty.
• No revenue, employee counts, decision makers or WhatsApp availability.

Runtime

The run uses its own APIFY_TOKEN only to start the dependencies and read their results. It is never sent to scraped websites. Three businesses are processed at a time, which fits the memory limit of a free Apify account including the dependency runs. Results and dependency run IDs are checkpointed, so a restarted run resumes without repeating finished work or charges. Budget about 20-25 minutes per 100 businesses.

Development

Python 3.13, Apify SDK 4.0.0. python -m unittest discover -s test covers identity rules, contact contamination, description filters, input limits, SSRF-safe URLs, verification semantics, billing and resume behavior. network.py and the email rules in extract.py are shared with Website Email Finder; keep security fixes synchronized.

---

Built by LeadProof - verified local-business lead lists, built to order.