800website builds WordPress automation tools, including an AI lead scraper that runs on the Google Places API. This is how that data source actually behaves.
What does the Google Places API give you?
Verified business records: name, phone, website, address, rating, review count and a Maps link. It is the same data behind Google Maps listings, which makes it considerably more reliable than scraping directory sites — the businesses exist, and the phone numbers were valid when the listing was last updated.
You query it with a text search — “dental clinics in Dubai Marina”, “accountants in Abu Dhabi” — and it returns matching places. Results come in pages of twenty, so three pages gives sixty records, which is a practical ceiling for a single scrape.
What it does not give you is email addresses. Google does not hold them. That single gap is why raw Places data is not a usable prospecting list on its own.
What does it cost?
Google charges per request, roughly $0.04 at current rates — about $0.40 for 200 leads. That is paid to Google directly against your own API key, not to a middleman.
Worth knowing before you budget: the cost is per request, not per usable lead. A search returning sixty businesses of which twelve fit your criteria still costs three requests. Tight, specific queries are cheaper per useful record than broad ones.
Set a billing cap on the API key. It is the single most useful five minutes you will spend, because a misconfigured loop against a paid API is an expensive way to learn.
How does AI enrichment fill the gaps?
The missing email is solvable because most businesses publish one — on their own website, where Google’s listing data does not reach. An AI model can read the site and extract the contact address, social profiles and a description.
That is what Gemini does in our scraper: takes the website URL from the Places record, reads the pages, and returns the contact details the listing lacked. It is not guessing at addresses; it is reading what the business published.
The limitation is honest: businesses with no website cannot be enriched this way, and some publish only a contact form. Expect a proportion of records to remain phone-only, and treat that as normal rather than a fault.
What about duplicates and regional filtering?
Two practical problems emerge as soon as you scrape more than once.
Duplicates. Search “restaurants Dubai Marina” and “restaurants Dubai” and you will get the same businesses twice. Deduplication needs more than one key — email, phone, name-plus-city, and Google’s own place ID together — because any single one of them misses cases.
Region. A search scoped to a city still returns businesses with international numbers. Filtering by phone prefix — +9715 for UAE mobile, for instance — is a blunt tool that works well.
Is this legal and appropriate?
Using the Places API within Google’s terms is a normal commercial use of a paid product. The judgement is about what you do next.
Business contact details published for the purpose of being contacted are a reasonable basis for a relevant approach. Bulk messaging everyone you scraped with something generic is how you get reported, blocked and blacklisted — and WhatsApp in particular enforces this. Check the marketing rules that apply to you, keep the outreach relevant, and make unsubscribing easy.
A short list contacted properly beats a large one contacted badly, and it costs less to build.
See our AI automation solutions, or talk to us about setting it up.
Talk to us
800website is a Dubai web design and development agency with published fixed prices. Send the details and you get a fixed scope and a fixed price back — no hourly billing. Contact us, message us on WhatsApp, or call +971 58 578 5380.
