Field notes from agentic implementation.

Practical notes from shipping closed-loop AI workflows. Exception handling, internal tools, dashboards, and the adoption work that decides whether any of it sticks.

I'm Zack Shields, an AI consultant based in Orlando. I own and operate businesses that run on the same systems I build for clients , a bar with automated inventory workflows, a short-term rental portfolio handling 700+ stays a year on AI-supported guest operations, and service companies running on n8n, custom internal tools, and voice agents. These notes are what that work actually looks like: what broke, what shipped, and what the team had to learn for it to last past the first month.

For deeper reference material on the models and platforms behind these builds, see Resources, or browse the service catalog if you want this shipped in your operation rather than reading about it.