Let’s slow down; make less

Work feels like it’s getting harder.

It’s getting harder because everything is moving faster, without getting clearer.

Decisions are accelerating. Content is multiplying. Bandwagons are being jumped on. Somewhere along the way, understanding stopped keeping up.

That’s what last year revealed to us.

AI is powerful. Everyone can see that. But that power without clarity creates new kinds of risk. Risk to trust. Risk to real productivity gains. Risk to shared understanding inside groups, from communities to organisations to nations.

3 points beyond theory and vendor promises

We addressed these issues in real environments, not hypotheticals.

In complex enterprise contexts, mistakes don’t just cost efficiency and revenue; they cost trust and brand equity. Together with our clients, we’ve been looking at how the latest technology could be used to advance audience goals and organisational objectives, while safeguarding long-term relationships.

For people to be able to apply AI in their day-to-day lives, they need support through the biggest transition in generations.

In our dialogue with real users during workshops with Rosenfeld, UXPressia, Web Summer Camp Croatia, World IA Day Zürich, and the LavaCon conference, across industries, three ideas kept resurfacing.

1. AI is only useful when it’s reliable

AI is an accelerant. At a certain speed, it’s suddenly no longer clear where judgment ends and automation begins.

We advocate for cautious AI. Teams need to move on from exploring basic AI capabilities, to defining what it should be allowed to influence, under what conditions, and how.

When using AI, you need to think, often harder than before, because of its ability to deliver the appearance of quality that hides underlying issues. AI reflects and amplifies what already exists, both the good and the broken.

To address this, we’ve been focusing on how AI can be used upstream in processes like research, design, operations, strategy development, and governance to improve how whole systems and lifecycles work downstream.

Value only appears when human context stays in the loop, and is fed to AI in a controlled and strategic way.

In practice, that meant working hands-on with teams to:

Use AI in user research, journey mapping, and content design, all grounded in real human data
Connect analytics to business value, not vanity metrics like clicks and production volumes
Design workflows, content, and knowledge models that support coordination at scale between humans, systems, and AIs

We shared this thinking through the OmnichannelX podcasts and our ongoing Truth Collapse work on AI, trust, and information integrity.

2. Trust doesn’t scale by accident

Everyone is overwhelmed. Too many messages, inputs, stressors, changes.

We’re in an environment where trust is essential at all levels of society and business.

Behind every experience people trust is an invisible structure doing the work, shaping meaning, enforcing consistency, and connecting decisions to outcomes.

Trust is built by delivering reliably.

Trust at enterprise scale is built through structure: clear content models, coherent taxonomies, new collaboration models, and metrics that connect design decisions to real outcomes; not clicks for the sake of clicks, or new for the sake of new, AI for the sake of AI.

This is where many organisations struggle. They invest in tools, but neglect the underlying human change management and content architectures that make consistency possible.

AI is easy to learn at a basic level, but real productivity still needs templates, processes, guidelines, and, yes, good old-fashioned training.

3. Information environments shape behaviour

We tend to talk about information systems as if they’re neutral.

But anyone who has watched a team argue over metrics knows the argument isn’t really about the numbers; it’s about what the system is encouraging people to care about. This effect doesn’t stop at the meeting room; at scale and across decades, the same dynamics fracture communities and shape whole cultures.

AI accelerates this effect. It gives the impression, especially when behind the scenes in recommendation or search algorithms, that it simply processes or retrieves information under our control. This is false.

An AI is an active agent that decides based on learned patterns what gets seen, amplified, and therefore, remembered.

When systems optimise for attention instead of understanding, the consequences aren’t technical. They’re cultural.

Through the Truth Collapse series, we explored what happens when intelligent systems have goals that are misaligned with human wellbeing. Over at OmnichannelX, we focused on what organisations can do to prevent drift in their internal communities, risking it spilling into customer-facing experiences and brand perception.

The information environments we design today are the ones future humans and machines will inherit.

Slowing down, producing less

We’re at a fork in the road.

One path leads to more automation, more content, more output; less trust, less clarity, less shared harmony. It’s faster. It’s louder. And it’s already crowded.

The other path is slower by design. It treats AI as part of the whole system, not a shortcut around it. It prioritises coherence over clicks, value over volume, and experiences people can rely on.

In 2026, we’re choosing the second path.

We want to help clients structure and machine-readable meaning to their systems. For our own internal work, we’ve set a specific goal: create less, reuse more, and organise better throughout the year.

Not louder. Not faster. Better.

If you’d like some help, let’s talk