AI automation
Describe your ideal customer. Get outbound-ready.
A natural-language prospecting demo: type who you want to reach, and get back a targeted lead list with a personalized first-touch email for every name, built on Apollo and OpenAI behind an n8n backend.
The problem
Selling an AI outbound system usually means describing it in the abstract: trust that the AI can turn a target description into real leads with real, personalized outreach. A prospect hearing that pitch has no way to check it without hiring first.
That same live proof was worth showing more than once, so it needed to run as a standing demo rather than a one-off build for a single sales call.
What we built
Type a plain-English target, '8 cybersecurity CEOs at mid-size companies in the US,' and the engine parses it into structured targeting with a live AI call, guards against an unusable request with a deterministic check rather than trusting an AI yes/no, sources matching people, writes a personalized fit rationale and first-touch email per lead with a second AI pass, and packages everything into a CRM-ready CSV.
The frontend streams each of the four stages live as it runs, then lands on a result screen with the lead count, the emails written, a per-lead 'why they fit' line, and the full outreach email ready to copy. The same demo that proved the pitch in one sales conversation got reused, unmodified, in the next.
The result
The pitch, provable live: describe a target, watch the leads and emails build in real time
4
stages streamed live: parse, source, write, package
1
n8n workflow behind the whole demo
2
sales conversations it's already proven the pitch in