Notes

Fewer topics. More judgement.

These pieces do not explain the model landscape. They focus on decisions that make an AI initiative defensible — or unnecessary.

  1. AI readiness: what to clarify before building

    A concise review of workflow, data, ownership, evaluation and operations before an AI idea becomes a system.

    Read
  2. RAG and permissions: access before relevance

    Why a knowledge system must constrain sources by identity and permission before generation.

    Read
  3. From prototype to an AI system others can operate

    An engineering workflow for small complete slices, testable change and a clean handover.

    Read