Field notes
Notes on dependable systems, practical automation, responsible AI use, and the lessons that appear when workflows have to hold up outside a demo.
2026-06-14
← Back to home
The token tax is real: plan AI costs before they sprawl
Agentic workflows can consume far more than a single chat request. This article explains how to separate predictable work from the few steps that need a model.
2026-06-14Schema markup won't get you cited by AI. Here's what does
Schema is hygiene, not a GEO strategy. What actually gets you cited by ChatGPT, Perplexity, and Google AI Overviews is entity consistency and sourced direct answers.
2026-06-14Which automations actually matter? Start with the failure mode
The most useful automations often solve ordinary but consequential problems: repeated entry, missed handoffs, unclear ownership, and failures that nobody sees.