Tarek Guesmi
Tarek Guesmi · Riyadh

Tarek Guesmi. I ran the operations before I automated them. At a club with fifteen staff and six thousand members, the systems I built gave back about three hours a day.

Start a projectLab
AutomationWebMarketingAutomationWebMarketingAutomationWebMarketingAutomationWebMarketing

Selected work

All entries →
Entry 005 · Web2026—06

ClubHub — A Sports Club Management Platform

The club logged everything by hand: paper subscriptions, roll-call attendance, and reports assembled at month end. I built a platform that manages subscriptions, records attendance by QR code, assigns staff tasks and generates reports — with a separate dashboard per club.

3
Hours saved daily
10–15%
Renewal revenue lift
6,000+
Subscribers over 3 years
15
Staff on the system
Status
Shipped
Disciplines
Web · Automation
Stack
React · Vite · Supabase · Recharts · Zod · jsPDF · Docker · Traefik
Role
Designed, built and deployed it — while running the club that uses it
Lesson

I built the system while running the club — so I knew where the hours actually went. Someone arriving from outside the operation cannot see that in a requirements meeting.

Entry 007 · Automation2026—06

WhatsApp Agent for a Children's Sports Club

A full automated responder for parent enquiries: pricing, schedules, trial-class booking and complaints — with no human in the loop. The agent reads and writes to live sheets, and alerts the owner silently on failure rather than showing the customer a technical error.

Status
Shipped
Disciplines
Automation
Stack
n8n · GPT-4o · Google Sheets · WAHA · Docker
Role
Built end to end — from the server to the conversation script
Lesson

The model has no notion of today's date by default — any logic depending on "today" or "tomorrow" fails silently unless the date is injected explicitly into the message context.

From the lab

All entries →
Entry 006 · Automation2026—06

Daily Email Assistant

Reads the day's mail, summarises it, and drafts replies — arriving on Telegram with buttons to approve or edit before anything is sent. The point is to keep the decision human and the labour automatic.

Status
In progress
Disciplines
Automation
Stack
n8n · Gmail API · Telegram Bot API · OpenAI
Role
A personal experiment — 21 nodes
Lesson

The easy part of automation is making it send. The hard part is making it stop and ask before it sends.

Entry 008 · Automation2026—06

Mass Broadcast Campaign — An Experiment I Shut Down

A broadcast system reading a list from a sheet and sending personalised messages at randomised intervals inside a safe time window. It worked technically — then WhatsApp restricted the number. I shut it down myself.

Status
Archived
Disciplines
Automation · Marketing
Stack
n8n · Telegram Bot API · WAHA · Google Sheets
Role
Designed it, built it, and killed it
Lesson

A system working technically does not make it safe to run. Broadcasting from the same number as the operational bot risked the whole asset for a short-term gain — separating the numbers is not an optimisation, it is a precondition.