12:01NEWThe 5 AI Automations I'd Build First (In This Order)
Effort, risk and control for each one — the order I'd build them again.
New video every Wednesday
The Suit WHO BUILDS
AI agents that run my inbox.
Built by a guy who ran companies for 15 years — and is not a developer. No hype, no vendor demos: real systems, real numbers and everything that broke.
- 5 episodes
- 1 every Wednesday
- 0 hype
$ git log --oneline a3f21c9 morning-brief OK 8c04e1d email: 3 drafts f19bb02 budget: clean 77d2a4e memory: +2 facts b5e90cf fix: silent cron 1e8aa37 synthesis: ship
● ran while I was asleep · nothing sent without me
$ latest --episode
The 5 AI automations I'd build first (in this order)
Five things ran this morning before I got up, and I didn't touch any of them. The five automations I'd build again — with the effort, the risk and the control for each one.
- 01The morning brief
- 02Email triage + drafts
- 03Meeting notes into tasks
- 04The monthly report
- 05The knowledge base
$ cat architecture.md
Run it like a company
Three layers, the same ones every well-run business already has. That's why you don't need to be a developer.
-
01
Knowledge — the filing room
Plain files the system reads every time: who I am, what I answer for, who the people are, how I write. One page, kept true. Every new chat stops being a new assistant who has never met you.
Episode: why it forgets you ↗ -
02
Rules — the SOPs
Written down, not hoped for. It drafts, I press send — that is not a setting. Every figure carries its source; conflicts get flagged, never silently resolved.
Episode: the 15-line email rules ↗ -
03
Agents — the staff
Scheduled runs do the grunt work overnight: the morning brief, email triage, meeting notes into tasks, the monthly report. I keep the judgment.
Episode: the 5 automations to build first ↗
$ ls ./episodes
All episodes
Newest first. Every one is a system I actually run — and what it cost me to get it right.
12:01NEW
9:00What Is an AI Agent? 3 Questions to Ask Before You Build One
Think about it like hiring. Three tests to run on any AI proposal before you say yes.
8:19Why You Need To Document Your Processes Now
Every new chat is shift change. One page it reads every time, and five rules that keep it working.
14:39Is AI Ready to Manage Your Email? I Tested It
The 15-line triage procedure, a live demo, and six ways it went wrong before it went right.
8:16START HEREI Built an AI Agent to Run My Life: Here Is What Broke
The architecture in plain business language, live demos — and the part most AI videos skip.
$ cat topics.md
What this channel is about
How to actually build with AI if you run things for a living — not if you write code for one.
AI agents, explained
What an agent actually is, what it's good and bad at, and the line it never crosses without you. Think about it like hiring.
Your inbox, your rules
Triage and drafts every morning. It drafts, I press send.
Memory that sticks
The one page your AI reads every time — so it stops forgetting who you are.
Building, no CS degree
Claude Code and plain files for people who define processes and check outputs — which, if you've run anything, you already do every day.
15years running companies
750people in the turnaround I led
0emails sent without my OK
1/wknew video, every Wednesday
$ whoami
Hi, I'm István — the suit who builds.
I started out designing engine parts at AVL and Bosch, then spent fifteen years running companies: a facility management business that grew from 50 to 350 people, group-level financing, and the turnaround of a 750-person company that reached break-even in under twelve months.
At 45 I stepped away and built an AI agent system instead — one that reads my email, runs my household budget and does my planning. In production, every day, since February 2026.
“Nobody sold me this. No SaaS subscription does it. I built it — and I'm not a developer.”
This channel shows executives and professionals how to actually build with AI: real systems, real numbers, and everything that broke along the way.
Come build with me
$ tell-me
What's the one task you'd fire yourself from tomorrow?
The comment section is where the next episode comes from. Tell me under the latest video — the best answers get built on camera.
e.g. chasing invoices every Friday$subscribe
no hype · real systems · real numbers