Field notes
Notes from the field
What we learn designing and deploying AI automations: what carries work, what stalls, and why.
AI Agents21 min read
AI Agents for Business: Where They Pay Off and What They Need First
An AI agent reads incoming work, picks from a fixed set of tools, and carries the task toward done within limits. Where agents pay off and how to pick one.
AI Automation7 min read
Intelligent Document Processing Without the Enterprise Suite
A document arrives and the automation reads it into a structured record with named fields. What that record carries, and where extraction goes wrong.
AI Automation7 min read
Human-in-the-Loop AI: Which Steps Wait for a Person
In a deployed automation every step either runs and reports or waits for a named person. Human-in-the-loop is the decision about where that gate sits.
AI Governance10 min read
EU AI Act Compliance Timeline: What to Stand Up Before December 2027
The EU AI Act's high-risk obligations apply from December 2027. The record they require is retrospective, so the program has to run in 2026.
Agentic AI6 min read
Autonomous AI Agents: Autonomy Is a Dial, Not a Switch
An autonomous AI agent carries work from trigger to done without a person starting each step. What autonomy means, and where it has to stop.
Agentic Workflow6 min read
What Is an Agentic Workflow? A Working Definition and a Field Test
An agentic workflow lets a model choose the next step instead of following a fixed sequence. A working definition, the anatomy, and a five-question test.
AI Automation5 min read
What Is AI Automation? The Definition We Build Against
AI automation is software that carries work from trigger to done, with a model handling the judgment steps. The definition, its boundaries, and examples.
Private LLM7 min read
The Private LLM Option: What Running AI on Your Own Hardware Actually Takes
Private LLMs are practical now. What a self-hosted model can and cannot do, what the hardware really costs, and when the API is still the right call.
AI Automation Tools5 min read
AI Automation Tools Are Not AI Automation
The tool list is the easy part. Why a stack of AI automation tools doesn't automate anything by itself, and what turns tools into work that runs.
AI Strategy7 min read
Why Your AI Needs Rules: Bridging the Gap Between Deterministic and Probabilistic Reasoning
The future of software is hybrid: deterministic rules around probabilistic cores. A guide for architects moving from rigid code to fluid patterns.
AI Strategy9 min read
Why AI Pilots Fail to Scale: The Work Is in the Handoff
AI pilots stall because they prove the model and skip the deployment. What the 95% figure measured, what 2026 data shows, and how to scale a pilot.
AI Readiness12 min read
AI Readiness Assessment: The 5 Dimensions Where Projects Fail
What an AI readiness assessment checks: five dimensions (data, technology, people, process, politics) and the questions to answer before the build starts.
Technical33 min read
Optimizing WSL2 for Claude Code: Complete Performance Tuning Guide (2026)
WSL2 setup for Claude Code, checked against current docs: quick start, .wslconfig sizing, sandbox, MCP, hooks, networking, and troubleshooting by symptom.