This week marked my anniversary at ZN. More than 20 years in, I’ve seen the same pattern repeat itself: only the technology changes.
There was a time when every organisation thought it needed a custom-built website. Then came content platforms, dashboards, monitoring tools. Each wave promised efficiency. Each wave delivered… more activity.
Today, it’s AI.
And once again, the instinct is the same: build something new, wrap it nicely, call it innovation. A new interface. A new dashboard. A new “assistant.” Another shell. And often, another license.
But the real question hasn’t changed: what problem are we actually solving?
In Brussels, we already operate in a kind of post-scarcity environment for communication. There is no shortage of content, newsletters, updates, or stakeholder engagement formats. If anything, there is too much. AI doesn’t fix that, it exposes it.
When every team can generate outputs instantly, the gap between activity and impact becomes impossible to ignore. More content doesn’t mean more influence. Faster production doesn’t mean better decisions.
At the same time, budgets are quietly shifting. Not necessarily towards better systems, but towards more tools. More subscriptions. More interfaces layered on top of each other. More shells: each with its own license, its own logic, its own limitations.
What looks like progress can quickly become fragmentation.
At ZN, we never positioned ourselves as builders of tools for the sake of it. Our role has always been to understand the client’s reality, map the constraints, and design the architecture that actually works for them. That used to mean choosing the right CMS, structuring content flows, defining governance. Today, it means something else, but also exactly the same.
It means asking where knowledge actually lives inside the organisation, how it is structured and maintained, what decisions need to be made and how often, and where human judgement is non-negotiable. Without clear answers to those questions, AI doesn’t create value, it simply amplifies noise.
Most organisations approach AI from the wrong starting point. They begin with outputs. A chatbot. An automated newsletter. A system to generate summaries. In other words, another shell, often sitting on top of a model they are already paying for elsewhere.
But outputs are the last step, not the first.
What actually creates value is what sits underneath: a reliable source of truth, structured in a way that can be used; workflows that define how decisions are made, validated, and repeated; and only then, outputs that reflect both. When those layers are missing, even the most advanced tools produce something that looks right but doesn’t lead anywhere.
There’s another uncomfortable truth. Much of what looks advanced in AI today has already been commoditised. The models are powerful, accessible, and increasingly similar. The advantage no longer sits in the tool itself, but in how it is used—and more importantly, in what it is connected to.
Paying for multiple shells doesn’t change that.
Without grounded data, without structure, without clarity on decision-making, even the best system will generate content that sounds convincing but has no real impact.
After more than two decades, that’s probably the one thing that hasn’t changed.
Technology evolves. The pressure to move faster increases. The tools become more powerful. But the organisations that actually create impact are still the ones that step back, design properly, and focus on decisions, not just outputs.
Not shells. Not licenses.
Decisions.
This article was originally posted by Jesús Azogue on Linkedin
