Writing on operational design, organizational knowledge, responsible AI, and the environments through which people and technology work together.
Agentic systems can be valuable, but autonomy is not appropriate for every task. A well-designed operational environment distinguishes among deterministic software, workflow automation, AI assistance, and bounded agentic work according to the nature of the task, the sensitivity of the information, and the consequences of error.
Models and infrastructure do not create a complete operational system. Organizations still need the layer connecting workflows, knowledge, people, interfaces, permissions, governance, documentation, and AI.
Organizations depend upon intellectual property, internal procedures, client information, research, archives, and accumulated expertise. They need meaningful choices about where that knowledge resides and which systems can access it.