A lab for studying how AI agents actually get used.

I’m building this lab as part of my broader research into practical AI adoption. The consumer narrative around agents is loud and often hype-driven. What’s missing is a grounded picture of where these tools genuinely help, and where they quietly fail.

I’m especially interested in how small businesses, creators, freelancers, and marketers actually use agentic workflows — not in a demo on a stage, but inside the messy, real tools they already work in every day.

The focus here is not hype. The focus is execution: what saves time, what improves quality, what fails, and what needs human judgment. Each agent in this lab is a small, honest experiment toward answering those questions.

I also consult on specific use cases — helping people design safe, useful workflows inside their own tools, with the right guardrails and a human firmly in the loop. The demos here are deliberately limited: you can test a use case, but deploying a real agent is something we scope together.