Insights, strategies, and perspectives on digital transformation, MACH architecture, and building exceptional customer experiences.
FeaturedThree days at Felix Meritis. Three AI Exchange hackathon builds shown live on the Technology Track, an open-source Agent Build Lab that invited the room to break it, and a community that has stopped asking whether agents work. What it proved, and what nobody on stage could prove yet.
20 articles
I spent two days at AGNTCon and MCPCon Europe in Amsterdam. Almost nobody on stage was worried about model capability. The blocker they kept describing was trust, and the teams that skipped a stage paid for it in months.
I spent an evening at AI House Amsterdam for the MCP Release Party. Here is what the 2026-07-28 specification actually changes for enterprises, and what it does not fix, from someone running ten agents in production.
The demo worked. The production system does not. So the team swaps in a newer model, rewrites the prompt, and runs the demo again. The model is rarely the problem. Nobody wrote down what good means, so nobody can prove this version is better than the last one.
The web is splitting into a human web and an agent web. Your expertise needs a front door on both, and most of the web is being locked out of the layer that names you.
You did not approve Grok Build. You did not procure it. It is not on your vendor list. None of that matters. One engineer installed it, pointed it at a repo, and the whole thing left the building, full git history included. Here is what was actually proven, what you are exposed to, and what to do in the next 72 hours.
Part three of three. The system could not improve itself, so I created a Director role and let one AI hire another to fill it. What happened next was not a technology problem. It was a management problem, and it taught me more about deploying AI than anything else I have built.
The EU pushed the high-risk deadline to December 2027 and a lot of executives exhaled. That is the mistake. The bans are already live, the transparency rules hit August 2026, and the delay did not touch either one.
Part two of three. How nine agents share one mind, why some of it lives in a database and the rest in a folder of plain text files, and what happens at two in the morning when the system quietly tidies its own memory.
The honest story of an autonomous agent ecosystem, built solo, on an 8GB box that used to be someone's office desktop. What it is, why it exists, and how nine AI agents got stood up in fourteen minutes. They did not, really. Keep reading.
Three days, more than 20,000 people, and one settled question. Not whether AI matters, but what to do with it, and what happens to the organizations that get it wrong. What we heard, in plain language.
I made three videos for queer founders this Pride. Here is what they have to do with starting this company.
Three AI models. One question about a portable AC unit. Three completely different answers. What it taught me about picking the right LLM, giving real context, and the one feature most people never use.
Most companies are told to clean their data before starting an AI project. That advice is expensive, slow, and wrong. Here's how to bake data cleanup into the AI rollout itself.
AI isn't the answer to everything. Learn when deterministic logic outperforms AI in your workflows, how to avoid AI runaway, and where intelligence actually earns its token cost.
Fidget Labs built a free interactive assessment that scores your composable maturity and AI readiness against 600 enterprise leaders. Here's the story behind it and what the data reveals.
A practical guide to avoiding expensive mistakes in AI adoption
If you're still trying to figure out the difference between a CMS, a DXP, and MACH architecture, or wondering why everyone keeps talking about "composable", you're in the right place. Let's trace the evolution of digital experience platforms from the early days of content management to the AI-powered future that's already knocking on your door.
Every enterprise leader dreams of the greenfield project. A blank canvas, no legacy baggage, pure innovation. Here's the uncomfortable truth: in enterprise technology, greenfield doesn't exist. But that's actually good news, if you know how to work with it.
Practical strategies for integrating AI tools into your daily workflows without losing the human touch
So you've decided to go composable. Congratulations! You're about to join a growing movement of enterprises breaking free from monolithic madness. But before you dive headfirst into headless, let's talk about the heartaches we've seen, and how to avoid becoming another cautionary tale.
Curious where your organization lands?
Take the free MACH & AI Readiness Assessment. Powered by the MACH Alliance Enterprise Technology Report 2026.
One short email when a new Focal Point article goes live. No spam, no sales pitches, just the writing.
Whether you're exploring MACH architecture, need composable strategy guidance, or want to discuss how we can accelerate your transformation journey, let's chat.