Workflow Automation

Self-Running Sales Revenue Engine

UNTH.AI (Internal Project) · Sales Operations

The Problem

Filling a sales list by hand is painfully slow: look up companies one by one, hunt for the right email, check it is real, and decide if they are even worth contacting — all before you can start selling.

What We Built

We built our own engine that does all of that on its own. It goes out and finds businesses, gets their contact details, works out which ones are a good fit, and even sends the first friendly emails — with a person giving the OK from their phone first.

How It Works

  1. It searches many websites and directories to find businesses that could be customers.
  2. It finds their email address and checks that it is real.
  3. It scores how good a match each business is, using AI.
  4. It writes and sends the first emails, once a person approves them from a chat app.

What Changed

The list fills itself with hundreds of thousands of accurate, ranked contacts, so the team can spend its time talking to the ones who reply instead of doing research.

CLIENT:UNTH.AI (Internal Project)
INDUSTRY:Sales Operations
TECH STACK:
PlaywrightHunter.io APIAnthropic Clauden8nBullMQPostgreSQLStripe

Interactive Workspace Sandbox

Explore how this system operates in real-time. Interact with the panels below to view simulated pipelines and data states.

Interactive Lead Scraper & CRM Pipeline

PIPELINE FLOW STAGES:
Crawl WebsitePENDING
Puppeteer retrieves page HTML & metadata
AI EnrichmentPENDING
GPT parses company size, industry, and value prop
Email DiscoveryPENDING
Hunter API scans for valid business email contacts
Lead ScoringPENDING
Evaluation model scores fit (High / Medium / Low)
CRM UploadPENDING
Syncs lead details directly to HubSpot database
🔍Enter domain and click Scrape Lead to view the structured pipeline results.

POWERED BY

Claude logoClaudeOpenAIn8n logon8nShopify logoShopifyReact logoReactCloudflare logoCloudflare

Systems, execution, results — built on the best AI stack.

Get Started