When your infrastructure vendor starts selling to your software
Cloudflare is aiming its network at software that buys and browses on its own. What that pattern tells you about the vendors in your own stack.
Read nowAI Leadership Journal
Digital Transformation
For leaders deciding how AI changes their organisation. The articles cover what to delegate to AI agents, how roles and tasks shift, what to build while the models keep changing, and how to run AI under regulatory limits.
Articles in this category
-
-
Managed agents and sandboxes: Anthropic vs OpenAI, for people who have to decide
Anthropic and OpenAI now host the agent runtime and its sandbox. How Claude Managed Agents and OpenAI's stack differ, and the token-vs-seat cost trade-off.
Read now -
The parts of an agent nobody draws for you
An agent harness is what surrounds the model: a loop, tools, a working memory, a trace. A field guide to the parts you deploy, trust and govern.
Read now -
Build an analysis agent that actually works
A worked build of an invoice-analysis agent: where analysis agents break, and how to make one you can trust to flag work for a human, not finish it.
Read now -
How to think about MCPs: what they are, why first attempts fail, and when to use one
The Model Context Protocol is the standard socket connecting AI models to your tools and data. What it does, what it doesn't, and when to use one.
Read now -
The job stays. The tasks move.
A Netflix leader's read on AI and the future of product and tech roles, and what a mid-sized company should actually do about it.
Read now -
What to build when the models keep changing
Each new AI model deletes another layer of the scaffolding built around it. A durable-versus-disposable test for where mid-sized companies should invest.
Read now -
An AI sandbox won't keep your documents secret. Buy one anyway
An AI sandbox bounds what an agent's code can install, delete, and reach. What it protects, what it never will, and how to buy one without overbuilding.
Read now -
Putting an AI agent in front of patients: the four disciplines that decide whether it's safe
A playbook for an AI agent in production in healthcare or any regulated business: bounded routing, deterministic gates, failure-driven testing, per-capability releases.
Read now -
The four questions every leader is really asking about AI
AI fluency for leaders is four leadership decisions, not a coding skill: delegation, description, discernment, diligence — the frame the whole column builds on.
Read now -
What to delegate to an AI agent (and what to keep in your own hands)
Delegation is the first AI-fluency skill: what to hand to an AI agent, how much autonomy to grant, and what a leader always keeps — plus one honest experiment.
Read now -
Running Claude Code on Amazon Bedrock for regulated data
Run Claude Code through Amazon Bedrock so regulated data stays on an approved path: a reproducible, verified setup for security-conscious European teams.
Read now -
AI Guide for German Mittelstand
A practical guide for German Mittelstand firms on AI adoption, data foundations, workforce readiness, and investment roadmaps to stay competitive.
Read now
