Hi, I’m Samuel 👋
I’m learning AI in public and sharing my journey with images, videos, prompts, and automation.
This is my first thread on what I’m learning and experimenting with. #AIinPublic#GenerativeAI
18 months ago, I took a course on AI Career Essentials and ever since then, my ambition to build the biggest AI community started.
Over a couple of months, the most tedious learning experience and character development 😂, I eventually became really good at using this tool called AI to automate stuff and create content.
After a while, I decided I might as well just kickstart that dream of building the community, so I birthed Flag Skool out of my Agency FlagIQ.
So there I was. Final-year EE student at UNILAG who just finished internship, with a laptop and a stupid amount of belief 🙂↕️
No money. No business plan. No clue if anyone would actually pay to learn from a 23-year-old.
But genuinely? I just wanted like 5 people I could speak my mind to and gist with about AI 😂😂😂
Here's what actually changed in all of that time, cos ngl a lot has happened:
- I stopped trying to be perfect and just started shipping. First "course" I taught had about 25 people. 7 of them ghosted by week 2 😂💔. I kept going anyway.
- I learned to charge what I'm worth. My first paying student paid me ₦100k. Standard pricing now is about 3x that 🌚. Same teacher, more belief.
- I figured out the actual gap. Most AI courses in Nigeria are either dollar-priced bootcamps nobody can afford or YouTube series that leave you stuck at the same wall. So I built the version I wished someone had built for me one taught by a friend, one I could relate to. Or at least if I watched it at my own time, I'd relate and learn faster.
- I stopped pretending I had it all figured out. Honestly mehn, half the time I'm still Googling things and roaming YouTube at 2am. Difference now is I have students Googling alongside me and Claude Code acting like a student on steroids 😂
Most of my last cohort started exactly where I started. No CS degree. No fancy connections.
Now they're landing gigs same as software developers, building cool stuff for themselves or their clients. All they had was curiosity and the willingness to actually build the thing.
If that's you, the next 12 weeks at Flag Skool could change a lot.
Cohort 2 details dropping 🙂↕️
Tbh I understand how content creation works but I think I’ll do it my way.
So for those that follow my page here’s how I’d be posting my content and I really hope you enjoy every bit of it.
I don’t know how to post daily. I’m the kind to pick up my phone, record a video, or type a tweet (like I’m doing rn) or just say something in the moment.
The whole having a structured post daily bores me, although I know it’s bad for my channel but at the end of the day I want to enjoy what I do do without having to make it a chore or work.
Anyways with that being said I’d appreciate if y’all just give me your honest reviews about my posts, engage and let’s build together.
On today’s showcase of an internal tool I built for my agent and we are using right now.
Here is Flag Tech. The a subdivision of my operation system FlagOS I am building.
Flag Tech is a management platform I built to keep track of how much work the devs in my agency put in when outsourcing technical expertise to clients plus how much revenue my agency is generating from the technical and software development aspect of the agency.
Why did I build this?
We were keeping records and data in scattered places, WhatsApp dms, Google Sheets and all of that but there was no unified space to see all my business operations. In other words I needed a dashboard, so yeah I built one.
Here’s the dashboard with dummy data.
Lmk what you think.
Somebody ordered Efo Riro, one Semo, one Pounded Yam, delivery to Ikeja Road, and not a single human being took that order.
That was Kylie. There's something about voice agents that's different from every other type of automation I've built, and I've been trying to put my finger on what it is. With a deterministic workflow, you define the logic, map every path, and the system follows it. Voice is different because the moment a human calls, none of that matters.
Humans interrupt mid-sentence; they change their order after confirming it; they ask something the knowledge base doesn't cover, and the agent has to handle all of it without the conversation feeling robotic.
The thing I kept coming back to while building Kylie was that sounding human is about how the conversation flows, the small acknowledgements, the way she confirms details before doing anything, the way she handles an unavailable time slot. Every one of those moments had to be deliberately designed.
Kylie answers calls, books reservations, takes food orders, sends confirmation emails, and logs everything, and the part I'm most proud of is that when you call her, she just sounds like someone doing their job.
Mimicking a human agent is straightforward; making something that feels like one is a completely different problem
just built an invoice chaser that runs itself 🧵
every morning at 9am it:
→ reads unpaid invoices from sheets
→ calculates days overdue
→ gemini AI writes the email
→ gmail sends it
→ slack tells me it fired
3 tones: gentle → firm → urgent
zero manual chasing. ever again
I delivered an AI web design project and got a great review 🔥
the shift is real — the best devs rn aren't just coders, they're directors
know what to build. know how to prompt. know how to make it *feel* premium.
#WebDesign#AIDesign#BuildInPublic