Find Business Contacts

Emails, phone numbers, WhatsApp contacts, social profiles - if a business publishes it publicly on its website or social media, DataMint’s AI can find it. This guide covers every finding tool: the Email Finder, Phone Number Finder, Deep Website Scrape, Facebook Page Scraping, and the playbook for hunting decision-makers (not just info@).

Found contacts → verify them before outreach: see Business Contacts Verification.

Email Finder

Run Email Finder (Have Website category) on your Website column:

1. DataMint crawls each site - the homepage first, then up to 2 contact-type subpages (/contact, /contact-us, /about, /team…) only if the homepage has no email.
2. It reads the fetched pages, extracting addresses from text, mailto: links, and rendered JavaScript content (browser-rendered fetch).
3. Placeholder and asset junk (template strings, image filenames) is filtered out.
4. Results land in a new column - hover a filled cell’s (i) icon to see what the finder actually saw.

Requirements: desktop app running + LLM API key.

What you get:

• Found emails per row, straight from public pages
• “None” when the site was checked but publishes nothing
• Empty cells when the run failed - retry those (Run Controls)
• Works on any website column: Maps scrapes, Epic Scraper imports, pasted data

Phone Number Finder

Finds phone and WhatsApp numbers from each site - more than just what’s printed in the footer:

Every number is normalized to E.164 international format (e.g. +8801712345678) and tagged by type - MOBILE or FIXED_LINE - so you know which numbers can even be on WhatsApp.

• Malformed numbers are dropped; numbers whose region can’t be determined are kept as unverified rather than thrown away.
• Region is guessed from the site’s content, the address country, or the domain’s country code (.bd, .co.uk…).

Requirements: desktop app running + LLM API key.

Deep Website Scrape

The deep option - extracts everything useful from each site in one pass:

How it fetches each site

1. Browser-rendered fetch (default): DataMint opens the site in your Chrome - JavaScript runs, then the fully rendered page is sent to the backend for extraction. This catches content that raw HTML never sees.
2. Static fetch fallback: if the browser fetch isn’t possible, the backend fetches the page server-side directly.
3. Sub-page crawl: if the homepage has no email, it checks up to 2 likely contact pages (/contact, /contact-us, /about, /team, …) - only when needed.

Requirements: desktop app running · no LLM key required (extraction is deterministic - no AI guessing) · source column must contain website URLs.

Live Status Checker (bonus)

Before spending API credits, check which sites are even alive. The simplest enrichment - and the only one that needs nothing (no backend, no keys):

• Sends an HTTP request to each website and records Active or Not Active (with the HTTP status code).
• Perfect first pass on a scraped list: kill dead-site rows first.

The Decision-Maker Playbook

Role addresses (info@, contact@) are what most businesses publish. For founder/owner contacts, combine three strategies - usually best in sequence.

Strategy 1: Finders + Deep Scrape (the base pass)

Start with extraction (above) - whatever is publicly listed lands in the columns.

Strategy 2: Google Search (Serper) for the person

When the site only has info@, search for the human:

1. Run Google Search on the business name with the founder template (default) - or email, linkedin, social, phone, or a custom query with {name} / {city} placeholders.
2. Up to 10 results per row, optionally summarized by the AI.
3. Follow up with the Facebook Scraper on any Facebook/LinkedIn link you find.

Requirement: backend + Serper key (+ LLM for the AI answer) - details: Google Search.

Strategy 3: Custom Prompt for pattern matching

The AI agent can reason about names + domains:

Find the owner or founder of {{Name}} from {{Website}}.
Look for their full name and any direct email address.
If no direct email exists but a name is found, suggest
the most likely pattern (e.g. first.last@domain) and mark
it as "pattern" instead of "verified".

Paste this into a Custom Prompt (mentions become {{Column}} chips) and run it on the Name + Website columns. See Use Custom Prompts.

Where Contacts Come From

DataMint only reads public data - what any visitor (or its AI) could see:

Troubleshooting

Related

• Business Contacts Verification - WhatsApp + email verification
• Google Map Scraper - build your lead list first
• Manage Data Table - how the columns fill in
• Cold Email Templates