Research focus
How small businesses, creators, freelancers, and marketers use AI agents — where they create value, and where they fail.
This lab is the practical arm of that research. Each agent is a small experiment; each demo is a chance to observe how a real workflow behaves when a real person tries to use it.
The questions
Where do agents create value?
Which tasks genuinely get faster or better — and by how much — when a small operator hands them to an agentic workflow.
Where do they fail?
The failure modes that matter in practice: brittle outputs, silent errors, over-automation, and tasks that look automatable but aren’t.
What needs human judgment?
The decisions that should stay with a person, and how to design workflows that keep a human meaningfully in the loop.
What’s actually worth building?
Separating the workflows worth investing in from the ones that are demos forever.
Who the lab studies
Small businesses
Operational repetition, customer communication, and the cost of context-switching.
Creators
Voice consistency, idea generation, and repurposing across channels without losing authenticity.
Freelancers
Positioning, outreach, and packaging — the business work around the craft.
Marketers
Angle generation, critique, and testing — speed without sacrificing quality.
How the demos fit in
The live demos are deliberately limited. They let people test a use case end to end without deploying anything — which keeps the experience honest and the observations clean. Nothing publishes, schedules, or stores data automatically. When someone wants to take a workflow further, that happens through a consult, where a real deployment is scoped for their own tools.