Most AI content is unshippable demos or jargon-stuffed think pieces. This kicks off a bi-monthly series that skips the theory and builds working GTM agent workflows step by step in n8n, Make, Claude, Gemini, and Cloudflare.
Full gist
The first build turns an inbound email into a full sales prospect brief in seconds, with copy-paste logic you can actually ship.
- published
- reading
- 2 min read
- original
- Originally on LinkedIn
On this page
… and start building it on Monday.
Most AI content right now falls into two buckets:
- Demos that took 3 weeks to build and will never hit production.
- Think pieces that use the word “paradigm” but don’t show a single line of logic.
A bi-monthly series showing you how to build actual agent workflows, step-by-step, using tools you can actually use (n8n, Make, Claude, Gemini, Mermaid, and Cloudflare).
No hype. No fluff. Just: Here is the problem → Here is the build → Here is the result.
But first… what does “Agentic” actually mean?
Forget the jargon. Here is the version that sticks:
An agent is software that doesn’t just respond - it acts.
| What it does | |
|---|---|
| Traditional Software | Waits for you to click a button or give an instruction. |
| Agentic Software | Receives a trigger, decides the “how,” calls its own tools, and delivers the “what.” |
The practical difference:
| What happens | Time | |
|---|---|---|
| Old Way | You get an email → research the company → write a brief → update the CRM → draft a reply. | 60 mins |
| Agentic Way | Email arrives → workflow fires → Claude (or any LLM) researches, synthesizes, and scores the lead → a brief hits your Slack before you’ve finished your coffee. | 0 mins |
The AI (Claude) is just the brain. The Agentic part is the nervous system - the orchestration that decides what happens next.
What we’re building (and what you can copy!):
I’m focusing on GTM motions, not ML engineering. We’re going to build:
Sales Prospect Intelligence: Inbound email to full brief in seconds.
Calendar Defense: A workflow that treats your deep work like a VIP.
Explainable Lead Scoring: Logic you can actually show your VP without a shrug.
The Enrichment Pipeline: Connecting messy data into a single scoring engine.
Reproducible. Copy-paste friendly. Built to ship.
Why now?
The gap between “AI is cool” and “AI is useful” isn’t a lack of power - it’s a lack of a map. The tools are ready. The APIs are cheap. You just need to see the steps.
First tutorial drops this week: From Inbound Email → Sales Prospect Intelligence. Built in ~2 hours.
Imagine every time an inbound email hits your inbox, a “digital researcher” has already done 60 minutes of prep work for you.
Before you even click “Open,” your Slack pings with:
Company Context: Latest funding, headcount, and mission based on open source data
Recent Signals: Did they just hire a new VP? Launch a new product?
The Pain Point: Why are they actually reaching out?
Suggested Talk Track: The 3 things you should mention in the first 5 minutes.
Confidence Score: Is this a Tier 1 lead or a tire-kicker?
The Reality: Most GTM teams spend 60+ minutes per prospect doing this manually. The Agentic Way: This workflow takes 0 minutes of your time once it’s live.
Gmail Prospect History - Agent analyzes all threads and creates a Profile to inform communication style.
Real data example - mining Gmail data to generate Behavioral Profile for the prospect:
Follow along if you’re tired of watching demos that don’t ship. Agent tech shouldn’t live in keynotes. It should live in your browser.
