AI
Artificial intelligence is quickly becoming part of the digital infrastructure behind modern websites and content systems. In this category, we explore how organizations are actually using AI in the real world, from AI-powered search and content workflows to governance frameworks that keep systems safe and accountable. The goal is simple: cut through the hype and focus on practical strategies that help digital teams adopt AI responsibly while improving the way websites, intranets, and digital services operate.
Learn More
Interested in saving time and having a secure website? Learn what Concrete CMS can do for you.
One of the hardest problems in building an AI assistant for a website or intranet isn't the AI part. It's the permissions part. How does the AI know what a specific employee is allowed to see? How do you prevent it from surfacing a confidential HR document to someone who has no business reading it? How do you keep AI responses in sync with content that changes constantly?
Every intranet has a graveyard. It's the section no one visits, full of pages that were accurate in 2021, linked to a tool that no longer exists, and last updated by someone who left the company eighteen months ago. Most organizations know it's there. Almost none have a reliable process for finding it before an employee does.
A supply-chain campaign called Miasma (attributed to TeamPCP) was surfaced on June 1, 2026, targeting npm developers. We don't use the affected packages in our work, but we ran a quick sweep anyway to be sure. This post describes what we checked and what we found.
If Google AI Overviews are getting in your way, here is the fastest fix I have found: add -AI to the end of any search query and the summary disappears.
We did not plan this. We did not run a campaign, hire an SEO agency, or engineer some clever technical workaround. We just looked at our search results one day and noticed that Concrete CMS was being cited in Google AI Overviews next to the University of California San Diego, Epic Games, and the Netherlands Cancer Institute.
Most people use AI writing tools the same way they'd use a search engine with extra steps. You type in what you want, you get something back, you sigh a little, and you start editing. The output is fine. It's just not quite right. The voice is off, or the details are generic, or it misses the actual point you were trying to make.
AI is no longer a pilot program. It is not something your forward-thinking employees are quietly experimenting with on the side. It is embedded in daily work, whether your organization has formally adopted it or not. The question is not whether your team is using AI. The question is whether they are using it consistently, safely, and in ways that actually move the needle.
Everyone wants to "use AI." Few organizations are actually ready.
That's not a criticism. It's just reality.
Let me say the obvious thing first: this post was drafted with Claude. That makes it either a very transparent piece of content marketing or a funny coincidence, depending on how you look at it.
I used ChatGPT for most of my writing work for the past few years Outlines, first drafts, turning long posts into short ones. It works. But a week ago I switched to Claude as my main writing tool, and I want to explain why.
Nobody's debating whether AI belongs in the military anymore. That ship sailed. AI is already embedded in logistics systems, training simulations, intelligence analysis, and even service delivery platforms that help soldiers find out where to get their car inspected on base.