Google Maps Lead Enricher (AI)

Picks up exactly where your Google Maps scraper stops. A Google Maps scraper gives you a pile of company names, phone numbers and websites. This Actor turns that pile into sales-ready leads: it visits each company's website, pulls the general contact email and social profiles, measures the website's quality, and uses OpenAI to score the lead and write a personalised opening line for your outreach.

You bring the list. It hands back a prioritised call sheet.

Quick start

1. Fill in the input — set the fields in the input tab (prefilled with a working example).
2. Click Start — the Actor runs and writes one record per item to the dataset.
3. Get your data — export to JSON, CSV or Excel, or pull it via the API.

---

The flow

1. Run a Google Maps scraper          2. Paste its datasetId here        3. Get finished leads
   (e.g. compass/crawler-google-   →      (or paste companies         →     emails · socials · web
    places)                                directly)                         signals · score · opener

1. Run any popular Google Maps scraper (for example compass/crawler-google-places) for your city + niche. Let it finish.
2. Copy its dataset ID and paste it into this Actor's datasetId field. That's the whole handoff — no reformatting.
3. Run this Actor. Each company comes back enriched and scored, sorted-ready for your CRM.

Prefer to skip the scraper? Paste companies straight into places instead.

---

What you get per company

For every company with a website, the Actor:

1. Crawls the homepage + up to 4 relevant subpages (contact, kontakt, about, om-oss, impressum, …). It uses a fast HTTP crawler first (CheerioCrawler) and only falls back to a real browser (PlaywrightCrawler) when a page needs JavaScript. robots.txt is respected.
2. Extracts contacts: general, role-based emails only (info@, kontakt@, hello@, office@, booking@, …), plus phone and links to Facebook / Instagram / LinkedIn / TikTok / YouTube.
3. Reads website signals: hasHttps, hasMobileViewport, hasOnlineBooking (known booking systems + "book"/"boka" wording), hasContactForm, cms (WordPress / Wix / Squarespace / Shopify / other), copyrightYear (from the footer), and a list of outdatedSignals.
4. If a company has no website at all → noWebsite: true. That is itself a strong lead, especially for web agencies.

Then the AI layer (OpenAI gpt-4o-mini, a fast and low-cost model) adds:

• leadScore (0–100) — how well the company matches your offer and how clear its gaps are.
• scoreReason — one or two sentences explaining the score.
• opener — a single personalised first line for a sales email, in your chosen outputLanguage, referencing one concrete, verifiable observation about that specific company. No invented facts.

---

Three ways people use it

• Web / digital agencies — find local businesses with no website, no HTTPS, no mobile layout, or a copyright year stuck in 2017. Lead straight into a redesign pitch.
• SaaS sales — filter a whole category (e.g. "no online booking", "still on Wix") to spot companies that fit your product, with an opener already written.
• Local marketers — build a ready-to-contact list with the right general inbox and the company's social handles, scored by how likely they are to say yes.

---

Input

Field	Type	Notes
datasetId	string	ID of a Google Maps scraper dataset. Reads title, website, phone, address, categoryName, totalScore, reviewsCount.
places	array	Alternative to datasetId: [{ name, website, phone, address, category }].
myOffer	string	Required. What you sell — drives the score and the opener.
outputLanguage	string	Opener + reason language. Default en.
generateOpener	boolean	Default true. Turn off to only collect emails/socials/signals (cheaper).
maxPlaces	integer	Cap how many companies to process. Default 100.
maxConcurrency	integer	Pages crawled in parallel. Default 10.
siteTimeoutSecs	integer	Per-site hard timeout. Default 40.

Provide either datasetId or places.

Minimal input example

{
  "datasetId": "aBcDeFgHiJkLmNoPq",
  "myOffer": "We build fast, modern websites with online booking for local service businesses.",
  "outputLanguage": "en"
}

---

Output example

{
  "name": "Sunrise Dental Clinic",
  "website": "https://sunrisedental.example",
  "phone": "+46 8 123 456",
  "address": "Main Street 1, Stockholm",
  "category": "Dentist",
  "totalScore": 4.6,
  "reviewsCount": 128,
  "noWebsite": false,
  "status": "enriched",
  "emails": ["info@sunrisedental.example"],
  "socials": {
    "facebook": "https://facebook.com/sunrisedental",
    "instagram": "https://instagram.com/sunrisedental"
  },
  "web": {
    "hasHttps": true,
    "hasMobileViewport": false,
    "hasOnlineBooking": false,
    "hasContactForm": true,
    "cms": "WordPress",
    "copyrightYear": 2019,
    "outdatedSignals": [
      "No mobile viewport (likely not mobile-friendly)",
      "Stale copyright year (2019)",
      "No online booking found"
    ],
    "pagesCrawled": 3
  },
  "leadScore": 82,
  "scoreReason": "Strong match: a busy, well-reviewed clinic whose site isn't mobile-friendly and has no online booking — exactly what your offer fixes.",
  "opener": "I had a look at sunrisedental.example and noticed patients can't book a time online yet — a gap worth closing for a clinic with 128 reviews."
}

A company with no website comes back with noWebsite: true, empty emails/web, and still gets an AI score and opener (missing web presence is a strong signal on its own).

---

Pay-per-event pricing

This Actor uses pay-per-event billing with two events:

Event	Charged when
lead-enriched	Once per company whose website was crawled (emails, socials, signals extracted).
ai-scored-lead	Once per company that received an AI score + opener. Priced higher to cover the LLM token cost.

The Actor logs its OpenAI token usage and the estimated AI cost per scored lead at the end of every run, so the event price always stays safely above the model cost.

---

Setup

The AI layer calls the OpenAI API. Set the environment variable / secret OPENAI_API_KEY on the Actor. Without it, crawling and extraction still run and ship — only leadScore, scoreReason and opener are left empty.

---

Use it from an AI agent (Apify MCP server)

This Actor is available through the Apify MCP server, so AI agents (Claude, ChatGPT, or anything that speaks MCP) can call it as a tool — hand it a datasetId and myOffer, and read back the enriched leads — with no custom integration code.

---

Run it locally

npm install
# put your input in storage/key_value_stores/default/INPUT.json
export OPENAI_API_KEY=sk-...   # optional, enables the AI layer
apify run -p

---

A note on data

This Actor only collects general, role-based company contact addresses (such as info@, kontakt@, booking@) published on a company's own website. It deliberately filters out addresses that look like they belong to a named individual.

Related actors

• Email Finder — Find email addresses for any domain — crawls the homepage plus contact/about/team pages
• AI Cold Email Personalizer — Turn a list of domains into personalized cold-email openers — AI reads each company's
• Google Reviews Insights (AI) — Turns Google reviews into concrete insights for a business and its competitors: top