Everything below runs on my own self-hosted n8n or was built and tested end to end. Where something is a capstone project rather than a client deployment, it says so. Client case studies are added here as agency builds ship.
The same kinds of systems I build for agencies: content engines, production pipelines with approval gates, research, and alerting.
LinkedIn AI Posting Agent
Problem: Consistent LinkedIn content requires daily research, writing, and publishing time most founders don't have.
Result: A Telegram controlled AI agent that researches, drafts in my voice, generates images, and publishes to LinkedIn. In chat it asks me twice before anything goes live. On its Monday to Saturday schedule, every draft must pass an automated voice check first, with a hard stop after five failed tries. Live, and checked against raw execution data.
Problem: Publishing consistently means research, drafting, formatting, builds and checks every week, and AI drafts can go live unreviewed.
Result: A weekly production run researches keywords, drafts, formats to the site template and runs a full build. Nothing publishes until I approve it by name. The approval gate exists because an early version published unreviewed.
A research analyst that never sleeps
Problem: Finding what's actually worth writing about means hours of manual research.
Result: Several sources pulled in parallel, AI-scored, and a ranked digest delivered to Telegram on demand. It feeds my content engine.
Failure alerts on every workflow
Problem: Automations break quietly - you find out a week later, after a lead is already gone.
Result: One shared error workflow that pings my phone when a run fails, wired into every production workflow. Every client build ships with the same.
Problem: Multi-branch service businesses find out about unhappy customers too late, and replies are slow and inconsistent.
Result: Built with my TS Academy capstone team: every piece of customer feedback is scored, routed by fixed rules, a reply is drafted for staff to approve, and the right branch manager gets an email and Telegram alert. Passed a 5 of 5 live demo run. It is a capstone project, not a client deployment, and the base for the review management system I build for agencies.
Before I built automations, I was a growth marketer. It is why the systems I build are aimed at pipeline and delivery, not automation for its own sake.
As Social Media Growth Marketer at Great Grace TV, I grew the channel from 3,000 subscribers into a content engine with 10.3M+ lifetime views, 67K+ subscribers, and 1.3M watch hours.
It wasn’t luck - it was a funnel. 73.7M impressions turned into 4.8M views turned into 755.5K watch hours. Every stage engineered: the thumbnail won the click, the hook held the view, the content earned the return.
Growth isn’t only YouTube. At Ayoken (Web3 NFT marketplace), I built Discord and Telegram communities from scratch, grew Twitter 50%, and drove community-led acquisition that contributed to a $1.4M pre-seed raise.
Want one of these for your agency?
Book a 30-minute intro call. We'll look at where your team's hours go and which system would pay for itself first.
