Every framework starts as an experiment.
AICurious is the learning lab behind the Human-AI Loop. Before an idea becomes a framework, product, or piece of advice at AIGal.io, it gets built and pressure-tested here.
Test before you advise. Every experiment starts as a hypothesis, not a hunch.
Collaboration you can feel
Not product demos. Playable demonstrations of what designed human-AI collaboration actually feels like.
The Simulation
SimulatorFive minutes inside what designed collaboration feels like, versus the bolt-it-on default. The Loop, in motion.
Open the simulator → Loop in Motion, the walkthrough ↗Triad in Slack
SimulatorHow one human orchestrates multiple AI collaborators, shown in the interface teams already live in.
Open the simulator → The 12-panel infographic ↗Strategist as Conductor
SimulatorAn animated scene of a leader conducting an AI-augmented delivery team, orchestrating the work rather than becoming its bottleneck.
Open the simulator →Experiments that became products
Some in alpha, some still prototypes.
Team Context Card
AlphaGives an AI teammate the context of how your team actually works, so collaboration starts from understanding instead of a blank slate.
View the product →Daily Fuel
AlphaA GLP-1 nutrition companion that adapts to how you actually eat, not how a chart says you should.
View the product →BurnoutBuddy
In developmentA burnout early-warning companion that notices the drift before you do, and helps you act on it early.
View the prototype →Opportunes
In developmentA personal opportunity operating system. Human decides, AI ranks. Every card is an opportune.
View the prototype →Under investigation
Where the testing goes next, and why it matters now.
ShopTuner
A test bed for what AI agents do to marketplace businesses. When an agent becomes the last click, discovery, attribution, and trust all change, and the human may never see what got filtered out. ShopTuner is where I plan to test defensibility in that world: how a brand stays visible when machines, not people, make the final choice.
Read the thinking behind it ↗The early experiments
Before the methodology had a name, there were custom GPTs. Early tests of AI as a genuine collaborator. 2024–2025.
Pep'r
An AI thought partner. The first test of whether AI could push back, not just answer.
Casi
A caregiving assistant, built from lived experience of managing care for family.
Sommit
A wine companion. A test of AI as a warm, opinionated guide in a domain of taste.