LeadEngine

Projects Automation & Pipelines Lead Gen & AI Voice Outreach

Search queries — hotels, restaurants and law firms queued by area and state, each one processed

Stack

Python, Selenium, n8n, Vapi

Kind

Lead Gen & AI Voice Outreach

Category

Automation & Pipelines

Status

Private product — demo on request

Source

Closed

Lead Engine

Outbound is two slow jobs bolted together — finding the businesses, then actually calling them.

Outbound is two slow jobs bolted together. First you find the businesses — search a term, open a result, copy the name, the number, the address, go back, repeat. Then someone has to actually ring all of them, knowing nothing about the business they just dialled.

Lead Engine takes a query — business type, area, state, how many — and works it all the way to a booked meeting. It searches Google Maps, scrapes every matching Google Business Profile, follows through to each business’s own website and scrapes that as well, and stores the lot as one record per lead. An n8n workflow wired to a Vapi voice agent then calls each business with a pitch assembled from that business’s own details, handles the conversation, books meetings and runs follow-ups.

One query — business type, area, state, limit — drives the whole thing. Selenium works Google Maps and scrapes every matching business profile, then follows through to the business’s own website and scrapes that too, so each lead arrives with real context rather than a phone number. An n8n workflow hands that record to a Vapi voice agent, which calls and pitches using the business’s own details, books the meeting, and runs the follow-ups.

The scraped businesses — 41 leads, 31 with a website, filtered by stage, call and follow-up

The problem

Outbound is two slow jobs bolted together — finding the businesses, then actually calling them.

Query the map

Business type, area, state and a limit — that’s the whole input.

One lead opened up — its own website scraped through from the Google listing, phone numbers redacted

Two-level scrape

Every matching Google Business Profile, then through to the business’s own website where it has one.

Built with

Python · Selenium · n8n · Vapi

Call logs — 41 completed calls and follow-up recalls, the scraped businesses redacted

Voice agent calls

An n8n workflow hands each lead to a Vapi voice agent, which pitches using that business’s own details.

Books and follows up

The agent takes the meeting booking and runs the follow-ups rather than handing back a list.

The n8n workflow behind the calling — context assembled, Vapi dialling, transcripts and bookings written back
JobGennieAI & Machine LearningKeep scrolling for the next fileNext →