Most organisations don’t set out to confuse people.

Communication teams aim for clarity, design teams work hard to maintain consistency, product teams optimise for usability, and technology teams focus on scale and resilience. Individually, these efforts are thoughtful and well-intentioned.

And yet, over time, something subtle begins to happen.

Messages shift slightly. Experiences feel just a little off. Decisions that once appeared aligned start to diverge in small but noticeable ways. The organisation still recognises itself, but only just.

This is drift. Not failure. Not negligence. Drift.

Drift rarely announces itself. It accumulates quietly through a series of reasonable decisions made in isolation. It emerges in the handover between teams, in the translation from strategy to delivery, and in the small optimisations made under pressure. None of these decisions are “wrong” on their own. Collectively, however, they begin to pull the organisation off-centre.

Drift as accumulation, not breakdown

Historically, drift was most visible in communication. Brand messages softened; tone varied between regions; guidance was interpreted differently across departments. These shifts were noticeable and, with effort, correctable.

Design systems helped address some of this by creating shared patterns. They reduced duplication and improved efficiency. But they also introduced new forms of drift. Components were reused outside their intended context. Patterns were applied without understanding the underlying rationale. Consistency was achieved, but sometimes at the expense of meaning.

Drift, in other words, did not disappear. It changed shape.

Why drift accelerates in system-led environments

Today, drift is accelerating again.

As organisations rely more heavily on automation and AI-mediated interfaces, decisions about how an organisation behaves are increasingly made by systems rather than people. Interfaces assemble dynamically. Content is generated or personalised in real time. Interactions are shaped by rules and models that operate continuously, often without human review.

This is not inherently a problem. In many cases, these systems are faster and more responsive than their predecessors. The issue arises when organisational intent is not explicitly defined in a way that systems can respect.

When intent is implicit, systems will optimise for whatever signals they are given: efficiency, completion, or the reduction of friction. These are vital metrics, but they are not the same as reassurance, trust, or judgement, particularly in moments of stress.

When intent is assumed rather than articulated

Most organisations believe their intent is understood. Values are articulated; principles are agreed. From a strategic perspective, the work appears complete.

But much of this intent lives in presentations, PDFs, and conversations rather than in forms that teams and systems can reliably act on under pressure. As work moves from strategy into delivery, intent is interpreted. As systems are introduced to accelerate that delivery, they fill the gaps with default behaviour.

None of this is careless. It is the natural outcome of complexity. But over time, drift becomes embedded. Language shifts slightly between channels. Interaction patterns feel “off.” Behaviour varies depending on which system is mediating the interaction, rather than the situation the person is in.

Where drift becomes visible

Drift rarely shows up in ideal conditions. It becomes visible in moments of stress.

Consider the experience of a major service disruption at a train station or an airport. In the past, this moment involved a staffed desk and human judgement. It was slow, but it allowed for a contextual response. Staff could prioritise vulnerable passengers or provide reassurance.

Today, that same moment is mediated through mobile notifications, kiosks, and AI chat interfaces. Technically, the transaction is resolved. But this moment is not purely transactional; it is emotionally charged.

If the organisation has not explicitly decided what matters most in these scenarios, as in what tone to adopt, or when automation should defer to human judgement, systems will default to what they know: Complete the task. Close the loop. Move on.

This is how drift becomes visible: not in everyday interactions, but at the moments that matter most.

The misdiagnosis of fragmentation

When organisations notice this fragmentation, the response is often tactical. Guidelines are refreshed; review processes are tightened.

These actions treat symptoms, not causes. Drift is rarely a tooling problem; it is an organisational one. It occurs when intent is assumed rather than articulated, when ownership is diffuse, and when systems evolve faster than governance. In complex organisations, coherence does not emerge by default. It has to be actively maintained.

This is where concepts like “Brand DNA” must become operational. They must answer the hard questions: What must remain true even when everything else changes? What signals reliability under stress? Where is flexibility acceptable, and where is it not?

Preventing drift rather than correcting it

Drift has always existed, but its consequences are now magnified by speed and scale. Once drift is encoded into a system, it propagates instantly. Correcting it later is far harder than preventing it in the first place.

Preventing drift does not mean exerting tighter control everywhere. It means being deliberate about what is fixed and what is adaptive. It means designing for scenarios, not just channels; for moments of stress, not just success.

Drift is not the result of poor intent. It is the cost of complexity without clarity. Organisations that recognise this early can design systems that adapt without losing their centre. Those that do not often discover drift only once trust has already begun to erode.