At 9:17 on a Tuesday morning, an employee needs a document for a meeting later that day.

She needs it translated, shortened, adapted for a different audience and checked for tone. Not long ago, this would have meant sending a request to several people and waiting for the work to come back.

Today, she opens an AI tool.

A few seconds later, she has something that looks remarkably close to what she needs.

And that is where the story really begins.

The most significant consequence of artificial intelligence at work may not be that machines can now produce documents. It may be that people have begun to change their expectations of what work should feel like.

Once you have experienced an answer in seconds, a first draft on demand or a translation produced while you are still thinking about the question, it becomes difficult to forget that possibility.

The technology changes first. Then behaviour changes. And eventually, expectations change.

For organisations, this may prove to be a more profound transformation than the introduction of any individual AI tool.

A different sense of time

For decades, the rhythm of corporate work was largely determined by processes.

Information had to be gathered. Someone had to write the document. Someone else reviewed it. A translation might follow. Then came formatting, validation and distribution.

The sequence was familiar because everyone understood the constraints.

Waiting was not necessarily perceived as a problem. It was simply how work was done.

AI disrupts that shared understanding.

An employee who can produce a first draft in seconds may no longer understand why obtaining an internal document should take two days. Someone who can translate a text themselves may question a process involving several rounds of exchanges. Someone who can ask an AI system to turn a long report into a concise briefing may wonder why the same information is difficult to find internally.

None of this necessarily comes from impatience. It comes from experience.

Technology has demonstrated that many tasks can now be performed differently. Employees naturally carry that experience into the workplace.

And this creates an expectation that organisations cannot easily reverse.

The employee may no longer wait for the organisation.

For most of the history of the modern workplace, organisations controlled access to the tools and information people needed to do their jobs. If you needed a document translated, you contacted the translation team. If you needed a presentation created, you contacted the relevant department. If you needed information, you searched the corporate systems or asked someone who knew where to find it.

AI changes the balance.

The employee can increasingly go directly from need to result without passing through the organisation's traditional channels.

If the organisation provides effective tools, this can be enormously empowering. If it does not, employees will improvise.

They will use public AI tools. They will create their own versions. They will translate their own documents. They will rewrite corporate information. They will build small, informal workflows around whatever technology helps them get the job done.

From the employee's perspective, this is rational.

From the organisation's perspective, it creates a new problem.

The ability to create information is becoming decentralised, while responsibility for that information remains centralised.

That tension has barely begun to be addressed.

The content paradox

The strange thing about generative AI is that it makes content abundant at precisely the moment when organisations need to become more certain about what content they can trust.

AI can produce remarkably convincing text.

But convincing is not the same as correct.

A translation can sound perfectly natural while using terminology that is wrong for the company. A rewritten document can be elegant while subtly changing its meaning. A summary can be accurate in tone while leaving out the one detail that matters.

And when hundreds or thousands of employees are using different systems, prompts and sources of information, the organisation can gradually lose something that was once taken for granted: a common version of the truth.

The problem is therefore no longer simply how to produce content.

It is how to know which content deserves to be trusted.

That question becomes even more difficult when AI-generated content is indistinguishable, at first glance, from content produced by a human expert.

For some employees, the difference will be obvious. They will know when a translation sounds wrong, when terminology is inconsistent or when a statement does not make sense.

For others, it will not.

That creates an uncomfortable asymmetry. The people most capable of identifying poor content may be the least likely to need it. Those with less expertise may be the most likely to accept a plausible answer as a reliable one.

The traditional content model was built partly around this problem. Reviewers, editors, translators and subject-matter experts existed not simply to improve language, but to provide a layer of institutional judgement.

AI does not eliminate the need for judgement.

It makes it harder to see where that judgement sits.

Who is responsible when everyone can produce?

This may become one of the most important questions of the AI workplace.

When content was produced through a relatively linear process, responsibility was easier to identify. There was a person who wrote it, someone who reviewed it, someone who translated it and, in many cases, someone who approved it.

The chain was visible.

AI makes it increasingly difficult to draw that line.

Who is responsible when an employee asks an AI system to produce a document? Who guarantees the accuracy of the information? Who ensures that terminology remains consistent? Who decides which information can be used? And who determines when human review is necessary?

These are not merely questions of governance.

They are questions about how work itself will be organised.

The temptation will be to create rules: approved tools, mandatory reviews, restricted access, human validation. Some of these controls will be necessary.

But excessive control carries its own risk.

If employees are given powerful tools outside the organisation while being offered slow or cumbersome alternatives inside it, the organisation may unintentionally encourage precisely the behaviour it is trying to prevent.

The answer cannot simply be to put the brakes on AI.

Nor can it be to let everyone work independently and hope that quality will take care of itself.

The challenge is somewhere between the two.

Organisations will need to decide where autonomy creates value, where expertise remains essential and where content carries enough risk to require explicit oversight.

That will vary enormously depending on the nature of the work.

A draft for an internal meeting is not a regulatory document. A translation of a marketing email is not a legal contract. A summary of a public report does not carry the same consequences as a piece of financial guidance.

The question is not whether a human should remain in the loop.

It is which human, at which point, for which type of content, and for what reason.

That is a much more difficult question.

A workplace built around new expectations

Something fundamental is changing in the way people experience information at work.

Once employees become accustomed to getting an answer in seconds, creating a document on demand, translating a message instantly or adapting content to a specific audience, these capabilities quickly stop feeling exceptional. They become part of what people naturally expect from their working environment.

The expectation is no longer simply to have access to information. It is to have the right information, when it is needed, in a form that is useful and relevant to the task at hand.

Information is expected to be accessible wherever people work, easy to understand, adaptable to different situations and increasingly personalised. At the same time, employees expect it to be secure and trustworthy.

None of this is particularly surprising. It is simply the consequence of new habits created by AI and digital tools.

The important point is that these habits do not stay outside the workplace. People bring them with them.

And organisations that continue to operate according to yesterday's processes may increasingly feel disconnected from the way their employees now work.

This is why the impact of AI on work cannot be measured only by productivity gains.

It is also changing the unwritten contract between employee and organisation.

Employees are beginning to assume that the tools around them should help remove unnecessary friction. They expect information to be accessible, adaptable and useful. They increasingly expect technology to support them rather than require them to navigate the limitations of the organisation's internal processes.

This does not mean that everything should happen automatically.

It means that every additional step needs to have a purpose that remains understandable in a world where technology can do so much more.

The end of waiting, or the beginning of trust?

Perhaps the most important change brought by AI is not that work can happen faster.

It is that the old boundaries between creating, finding, adapting and consuming information are beginning to disappear.

The employee who once consumed corporate content can now create it. The person who once requested a translation can now produce one. The manager who once waited for an analysis can now ask a model to generate one.

The distinction between producer and consumer is becoming blurred.

That is enormously powerful.

It is also potentially destabilising.

When everyone can produce content, the value of production itself decreases. What becomes more valuable is the ability to determine whether the result is accurate, appropriate, consistent and trustworthy.

In other words, AI may not make content the scarce resource.

It may make trust the scarce resource.

And that could be the real transformation taking place in the workplace: not the disappearance of human work, but a gradual shift in what humans are expected to contribute.

Less time producing what machines can already produce.

More responsibility for deciding what should be produced, what can be trusted and what ultimately deserves to be acted upon.

The organisations that understand this distinction will be better prepared for what comes next.

Not because they will necessarily use more AI.

But because they will understand that once people have changed the way they experience information, they have changed the way they experience work itself.

‍