A few years ago, writing a difficult email or text took time. You stared at the screen, rewrote the sentence and worried about the tone, how it would land. Sometimes you delayed sending it because you were not ready to address the response that might come back.
Now, most leaders can use AI to draft that same message in minutes, whether they need help brainstorming or cleaning up the message. For routine communication, this is a real improvement. The words are clearer, the structure is sound, and it takes a fraction of the time to complete.
And still, there are messages you can’t rely on AI for. Feedback is one of them. Your AI-drafted message may be technically correct, but it does not always land the right way. It reads right, but feels wrong.
Feedback can be emotional, but that’s why companies have leaders to manage these tough moments. Leadership is about presence, not polish.
When AI Moves from Clarifying to Cushioning
At first, leaders feel AI’s impact in small, practical ways. Quarterly goals are written faster, planning documents feel more structured, and the blank page is less intimidating. It helps them ask better questions about outcomes, success metrics, obstacles and accountability.
These capabilities can strengthen an organization’s business operating system (BOS) by making communication clearer and execution more consistent. But the best operating systems still depend on leaders doing the hard, human work.
Not every moment can be standardized or templated. For example, a piece of feedback that feels uncomfortable or a decision that might disappoint someone. While it’s tempting to pass these off to AI, it’s going to hurt your company culture.
A study published in the International Journal of Business Communication shows employees can tell when a manager uses AI. Leaders who relied heavily on AI-written messages were rated as less sincere, less caring and less trustworthy.
Even though the words were correct, people felt something was missing. They were actually reacting to the absence of the person behind the message.
Imagine a manager sending a message about performance. The tone is calm, the language is thoughtful and the structure is solid. When the employee reads it, they think, “This sounds right, but it doesn’t feel like it came from my manager.” They sense distance.
In moments like this, people are absorbing information and looking for signals. Do you mean this? Are you willing to stand behind it? Would you say this the same way if we were sitting across from each other?
An AI solution can’t answer those questions.
Where Culture is Really Built
Culture is built in the small daily rhythms of work. All-hands meetings, strategy decks and company values play a role, but the real progress happens during personal interactions.
Culture lives in conversations where something is at stake: A performance review that affects someone’s confidence, a disagreement about priorities or a moment when expectations are unclear and tension rises.
How does a manager handle missed expectations? Is feedback direct or avoided? Are difficult topics addressed early or left to linger?
Think about which moments shape work experience more: A senior leader talking about accountability to the entire company or a manager sitting down with you to talk honestly about a specific missed deadline and concrete steps you can take to keep it from happening again. The second conversation is more uncomfortable but also far more influential. These are the moments where trust is either reinforced or weakened.
When leaders avoid hard conversations, teams notice. When misaligned behavior or performance issues go unaddressed, people take it as a sign that those things don’t really matter to their leaders or the organization as a whole. Over time, that avoidance becomes part of the culture.
In tough situations, AI can help leaders think through what they want to say or organize their thoughts before a conversation. But it can’t do the emotional work of leadership. It can’t sit across from someone, acknowledge tension, and say, “This is uncomfortable, but I want to help you improve.”
Employees are not looking for flawless scripts, but rather leaders who are present, honest and willing to stay in the conversation. A manager who says, “I do not have all the answers, but we can figure this out together,” builds more trust than one who delivers a perfectly polished message and then disappears.
This is where leadership becomes visible.
Leadership Skills Can Fade
Skills like honesty, empathy, listening and the ability to sit in discomfort are not static traits. They develop through use. When leaders regularly buffer hard conversations with tools like AI, those skills get less practice. This results in shallower leadership.
Leaders recognize this and say that the hardest part of AI adoption is not the technology. Instead, 93% said the biggest challenges are human issues like culture and change management. AI clearly works, but tension exists around whether leaders are willing to do the human work that is still required.
Trust Still Depends on Leaders
The way leaders talk about AI sets the tone for how people experience and adopt it. Leaders who build trust talk openly about it and experiment alongside their teams instead of simply mandating use. They also clearly distinguish AI as support, not surveillance.
They understand that people react differently. Some see possibility while others see risk. Some are concerned about structure, while others care most about relationships. Effective leaders adjust their message to each person. That’s something AI can’t do for you.
McKinsey’s State of AI 2025 report shows that 88% of companies use AI regularly in at least one business function, and 62% are experimenting with AI agents. AI is now part of everyday work.
As employees adjust to this new reality, they are looking for more reassurance. They want to know that someone is paying attention, that decisions are owned and that they can speak honestly and be heard.
At the end of the day, AI can help leaders strategize and communicate more clearly, but trust forms when they show up in real time, especially when conversations are hard.
That part of leadership cannot be automated, and most people can tell the difference.
