Build your first working AI system in 90 minutes.
A live beginner-friendly session where you turn one painful task into a repeatable workflow you can run at any time.
You build alongside me and overcome your fear of AI.
Register nowRecording included
A live beginner-friendly session where you turn one painful task into a repeatable workflow you can run at any time.
You build alongside me and overcome your fear of AI.
Register nowApril 30, 2026 · 1PM SGT (GMT+8) · Online via Google Meets
Pillar 1 · 0:00–0:30
Most people struggle to use AI effectively because they don't know what problem to solve with it. We go through a mindset shift of using AI to solve one painful bottleneck in your workflow.
Pillar 2 · 0:30–1:00
AI can finally work with your context instead of forgetting who you are or what you do. Learn how to install the key components and route different tasks to models for token efficiency.
Pillar 3 · 1:00–1:30
Outsource the soul-sucking tasks you hate doing to AI and build highly specific and personalised workflows so the LLM gives you high-quality outputs that are specific to your context.
I was an AI skeptic. I just saw it as a glorified search engine that gave me average outputs when I asked it to write tweets for me.
When I was on hospitalisation leave in Jan 2026, I set a goal to get good with AI. I was using it every day and doing exactly what you're probably doing: opening Claude/ChatGPT, re-explaining myself, getting mediocre outputs, and getting frustruated. I was wasting time re-explaining myself until I finally realised how AI can work for me.
I had to build a system that includes AI inside, and then outsource all the tasks that I hate to it with Agent Skills.
Signal Starterer gives you a full guide on how to get started with building your sysatem that works for you with the simplest tools possible.
Paste a YouTube link and get a full transcript and summary without doing all the steps manually.
2 min saved per video
Pulls transaction history from any DeFi wallet I follow and returns a plain-English summary of what they did.
3 min saved per wallet per day
Analyses and makes edits to my article's formatting and Twitter username mentions for consistency.
15 min saved per article
Scans my reply history and adds them to a database for voice analysis.
1 min saved per reply
A full course to run your life with AI.
Build a working AI system in a 90-minute live session
Signal Starter + a dedicated 1:1 session.
If you don't achieve a working Skill by the end of the 90-minute session, I will roll you over to the next session at zero cost.
Questions? DM @gideonfip on X.
I'm 100% confident that there are tasks in your workflow that can be automated with AI. Every boring, repetitive, soul-sucking task that you do day in, day out. All of them can be easily automated with AI.
So many tools are now integrating closely with the models, and it's easier than ever to build a simple workflow to get work done, even when syncing across multiple platforms.
It's possible to start this system with a free OpenRouter API key, but you will be subject to rate limits and the outputs may not be as effective as a subscription.
I will teach you how to optimise your model routing requests so you only use your subscription for thinking tasks and skill creation, while the free models are called up to execute workflows that don't require significant compute power.
No, I built out my system from scratch without knowing much about coding. You just need to know the outcome and what you want to achieve (i.e. the what), and AI will focus on getting the task done (i.e. the how).
This won't be something that you perfectly get in one go. I've consistently iterated on my system before it gives me the outputs that are satisfactory to me. But once you've created the workflow, you just need to give Claude the slash command, and it'll carry out the exact steps for you.
It's completely possible for you to use Claude Cowork if you're fine with paying for the Claude subscription. However, not many people are willing to pay that amount, and this session is designed to be flexible enough so you can integrate or add different models.
I'll teach you how to build an open system that can use any model and still get the same results, instead of being locked into the Claude ecosystem. There are cheaper LLMs out there with comparable outputs to Claude and higher rate limits.
While Claude is great and has many innovations that make it easier to use, there are also other cost-effective options you can consider to build out your system, instead of being forced to stick to one model.
No, you do not need to copy my setup exactly. The beauty of Signal OS is how you're able to build out a system that works for you using workflows that you are familiar with.
Everyone has different ways of writing and carrying out the same task, and what works for me won't necessarily work for you.
This session is meant to equip you with the skills required to build out your own workflows and your own system that works for you, instead of copying someone else's template.
Yes, every attendee gets the full recording, and the link will be sent within 24 hours of the live session. The recording follows the same build sequence — you can pause, rewind, and build alongside it. I'd highly recommend attending the live session if possible, so you can ask any questions live if you face any difficulties.