When Everything Sounds Intelligent, Who Is Doing the Thinking?

A few months ago, I read a piece of writing that was clear, confident and beautifully arranged.

It answered every point in the brief. The sentences flowed. There was even a thoughtful conclusion.

But something bothered me.

I could not tell what the writer actually thought.

This is becoming a familiar feeling at work. A proposal arrives quickly. An email sounds diplomatic. A strategy uses exactly the right language. Everything appears ready to send.

And sometimes it is good work.

Sometimes, though, the polish has arrived before the thinking.

I use AI. It helps me sort through information, test an idea and find a clearer way to express something I already mean. When a deadline is close and the page is empty, having somewhere to begin can be a relief.

I can also see why a young employee would turn to it before asking a manager a question. The tool is available at midnight. It is patient. It does not sigh when you ask it to try again.

The concern is not that people use it. I worry about the moment when a convincing answer makes us stop asking whether it is the right answer.

In communications work, I have spent many years looking for the detail a polished draft can miss. What happened before this announcement? Who might hear the message differently? Is the promise something the organisation can actually deliver?

A sentence can be grammatically perfect and still make a difficult relationship worse.

Imagine a company has disappointed a long-standing partner. Someone asks AI to draft an apology. It produces a courteous note that acknowledges the inconvenience and looks forward to continued collaboration.

The note sounds professional.

But the partner is still waiting for someone to explain what went wrong, who will fix it and when.

The problem was never a shortage of courteous words. It was a decision the organisation had yet to make.

AI can help write the message. It cannot take responsibility for what the message promises.

That distinction matters in more places than communications. It matters when a manager reviews a job applicant, when a team recommends a new market, and when someone prepares a report about people whose experience cannot be reduced to a neat summary.

Good judgment often begins by noticing what is missing.

Sometimes the missing thing is a fact. Sometimes it is local knowledge, the history of a relationship, or the courage to say that the brief itself needs to change.

This is also why experience matters. An experienced person may recognise that a proposal which looks sensible on paper will fail in practice. They have seen a similar decision before. They remember the complaint that came six months later, or the colleague who quietly kept the whole arrangement working.

A newer colleague has not had those years to build the same instinct. If AI supplies a finished answer before they have worked through the problem, where will that instinct come from?

Recent Harvard Business Review articles raise a related concern: AI makes polished work easier to produce, while people still need to learn how to judge what is worth trusting, questioning and changing. One article asks how less experienced workers will develop that judgment when the tool can do so much of the early work for them. [hbr.org]

That question stayed with me because I think about the people I mentor.

I want them to have useful tools. I also want them to experience the untidy part of learning: making a first interpretation, explaining why they chose it, hearing a different view and discovering what they overlooked.

There is a kind of confidence that comes from producing an answer quickly. There is another kind that comes from understanding how you reached it.

We need the second kind when the situation changes.

A person who has only learned to present a recommendation may struggle when someone challenges it. A person who has learned to examine the evidence can say, “You may be right. Let me look at that again.”

To me, that is an intelligent response.

Yet many workplaces reward the appearance of certainty. We ask for concise recommendations, clean slides and immediate replies. We praise someone who seems to have everything under control. Then we introduce a tool that can produce all of those things in seconds.

Perhaps we should not be surprised when fewer people show us the questions they are still wrestling with.

I have caught myself enjoying the speed too. There are days when a draft comes back so smoothly that I want to accept it and move on. Reading it properly takes more time. Checking a claim takes more time. Deciding that an elegant paragraph does not sound like me takes more time.

The machine has removed some work from the writing.

It has not removed my responsibility for what I put my name on.

In my August note, I wrote about what happens when an organisation’s purpose becomes a poster: the words remain on the wall, while the daily decisions tell another story.

I wonder if AI could create a similar gap.

We may say we value curiosity, originality and thoughtful leadership. Then we measure people by how fast they produce something that looks finished. We may say we want employees to think, while leaving them no time to think before the next deadline.

The result could be an office full of intelligent language and very little conversation about what anyone believes.

That would be a loss. Not because every piece of work must begin from a blank page. Most good work does not. We learn from other people, borrow structures, ask for advice and revise our opinions.

AI is another source of help. The question is whether we remain present in the work.

When I read something written with AI, I do not need to know which sentence came from the tool. I want to know whether the person who submitted it can tell me why the recommendation makes sense.

What did they check? What might be wrong? Whose perspective is absent? What would change their mind?

These questions should not be used to embarrass a junior employee. A manager can make it easier to think aloud by asking them with genuine interest. The answer may be incomplete. That is where a useful conversation begins.

Where I Would Begin

I would begin before opening the tool.

Write down the problem in ordinary language. Who is affected? What decision has to be made? What do we know, and what are we assuming?

Then use AI to help explore it. Ask for alternative explanations. Ask what the proposal may have overlooked. Ask it to challenge a first idea. Check the facts it gives you, especially when they will shape a decision or appear under your name.

And before sharing the finished work, put the tool aside for a moment.

Can you explain the recommendation to another person without reading from the page? Does it fit what you know about the people involved? Is there a sentence you would hesitate to say to them face to face?

If there is, pay attention to that hesitation.

For leaders, I would add one more step: give people room to explain their thinking before correcting the output. Ask how they arrived at a conclusion. Tell them what you noticed that they may not yet see. Let them watch you change your own mind when new information deserves it.

Judgment grows through that exchange. It is difficult to learn from a final draft alone.

I do not know exactly how AI will change every profession. I suspect some parts of our work will become easier, and others will ask more of us. Harvard Business Review has also warned that careless use can weaken the particular skills and knowledge that make an organisation valuable. That is a risk worth examining in our own workplaces. hbr.org

For now, I keep returning to a simple question when I read an impressive answer.

Who has taken the time to understand it?

The answer may have begun with AI. It may have been improved by a colleague, challenged by a customer and rewritten several times. None of that troubles me.

I would just like to know that somewhere along the way, a person paused.

They looked beyond the fluent sentences, considered the people who would live with the decision, and decided what they could honestly stand behind.

The page may sound intelligent either way.

The difference becomes clear when someone asks, “Why?”

—end—

 

About Yoke Darmawan

Yoke Darmawan is the founder of D&A Consultancy and an Indonesian strategic communications, brand and organisational engagement consultant based in Bali. Drawing on more than twenty-seven years of professional experience, she writes about workplace culture, leadership, communication, purpose and the human behaviour behind organisations.

Yoke’s Midnight Notes is her monthly collection of reflections on work, people and staying human. The subjects include workplace culture, leadership, communication, professional manners, employee engagement, brand reputation, partnerships, purpose, empathy and the changing relationship between people and technology.

Where useful, the reflections are supported by credible research and contemporary management thinking. They are observations from experience, written for people at every stage of working life. Yoke is not writing because she believes she has all the answers. She is writing because, after all these years, she is still paying attention.