AI can write your emails, but can it read the room?

Human judgement and emotional intelligence are crucial for leadership communication in the age of AI, write UNSW Business School's Gabi Nudelman and Kelsey Burton

In 21st-century workplaces where communication is constant, it’s no wonder we’re searching for better ways not just to keep up, but to be heard. 

After all, Microsoft’s latest 2025 report shows that employees are interrupted from their work every two minutes, racking up 275 pings a day from meetings, Teams chats, and email alerts. Nearly half of employees feel their work is chaotic and fragmented, and 80% say they lack the time or energy to do their jobs. Even off-the-clock time isn’t safe, with 56% of Australians feeling obliged to check work apps after hours, despite “right-to-disconnect” laws taking effect in August 2025.  

To manage this chaos, many are turning to AI tools that promise relief by automating business communication, but at what cost? When handing over business communication to AI, what do we lose? 

While AI can write your email, it still can’t read the room. AI can’t sense when to refrain from sending that email at 10 p.m., nor can it perceive a client’s anxiety behind a one-line brief. As we rush to automate, there’s a real risk that we scale poor communication habits, and that’s more than just bad practice - it’s bad for business. 

Why the human touch still matters 

Communication in business isn’t only about content; it’s about context, timing, and tone. Predictive text can mimic your writing style, but it can’t read between the lines of its coded text. Without being prompted or provided with contextual artifacts, the AI-generated, perfectly written communication can miss the mark completely.

Aspects of effective written communication that AI lacks are skills that even humans struggle with, such as:  

1. Reading undercurrents, such as cultural norms, team dynamics, and the “rules” no one says out loud.

2. Weighing power dynamics, like recognising when phrases carry unintended connotations (“as per our last conversation”) or being aware of our own perceptual limitations and biases.

3. Judging situational appropriateness, including understanding stakeholder perceptions, employee morale during times of change, and the right timing and tone for each communication.

Poor communication costs companies an estimated US$12,506 per employee, per year, based on an average salary of US$66,967. In professional environments, especially in leadership or cross-functional roles, tone, timing, and relational nuance are just as important as the content itself. The same words can reassure or alienate, inspire or irritate, depending on how they’re delivered.

While AI can help us learn to communicate effectively if prompted to do so, relying on it blindly risks reinforcing the very habits we need to unlearn. Because, in the end, it’s not just what you say, but how your message lands.

Leadership requires effective human communication  

In a leadership role, communication isn’t just about passing on information. It’s how you shape culture, signal priorities, build trust, and steady your team in times of ambiguity. 

When things are going smoothly, communication can be routine. But during change, crisis, or conflict, it becomes high stakes. A poorly handled message, even one that’s technically correct, can cause disengagement, confusion, or reputational damage. Tone-deaf announcements, abrupt changes with no context, or coldly delivered decisions can erode morale faster than the decision itself. 

Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School

One high-profile example came when Duolingo announced its move to an “AI-first” model in April 2025, replacing many contract workers with GenAI. While the company offered upskilling opportunities, critics noted the ambiguity around the CEO’s messaging and lack of empathetic communication. For those affected, the message felt more efficient than human. 

That’s the difference. Leadership communication isn’t just about clarity; it’s about care. It means understanding how a message will land, not just how it’s written. 

Leaders don’t just share information; they provide context and reassurance, set the tone and model values. When done correctly, this can have a measurable impact: organisations with strong internal communication see up to 45% higher employee engagement and a 32% boost in profitability.  

That’s something no chatbot can do for you. 

Learn to co-write with AI  

There’s no question that GenAI is a powerful tool. It can speed up drafting, sharpen clarity, and help organise messy ideas into neat structure. For busy professionals juggling inboxes and back-to-back meetings, that’s no small win. 

But there’s a risk in getting too comfortable. When we start relying on AI to do the thinking, not just the typing, communication becomes less intentional and more generic. The goal shouldn’t be to outsource communication. It should be to augment it. That means using AI to handle the mechanics, while reserving the judgment, tone, and audience awareness for yourself. 

A recent example comes from IBM, which replaced hundreds of back-office HR roles with AI systems, while simultaneously ramping up hiring in human-centric areas like sales and programming. The message is clear: AI can take over repetitive tasks, but the parts of work that require empathy, influence, and nuance are more valuable than ever. 

So, use AI to help you get started. Let it tidy up the structure or offer a first pass. But always bring your own lens, because your audience still expects a human behind the message. 

Lead with AI-aware communication practices  

Here’s how to keep your communication human, even when AI is part of the process: 

1. Start with intent, not the prompt. First, ask yourself, what am I trying to achieve with this message? How do I want the recipient to feel? Don’t let AI define your message intent or purpose. Take ownership of what you want to say, how you want to say it, and guide the AI tools as you co-construct your message.  

2. Automate mechanics, own the moment: AI will perfect your structure and tone. It will reorganise slides, analyse data, and develop reports. However, when writing performance feedback, crisis notes, or investor updates, we need to take ownership of the written communication we send. These are times when judgement and leadership are most critical.  

3. Embed local norms and empathise: AI is not aware of your colleagues’ situations beyond what you tell it. Sense-check whether your message might unintentionally dismiss or inflame. When developing communication for AI tools and support, teach agents about Australia’s right-to-disconnect hours and culturally preferred salutations. Default US templates often ring hollow here. 

4. Personalisation is key: At the end of the day, what keeps people around is belonging and connection. Ensure your voice is not lost with AI-generated communication. Refer to a shared moment or ask about life outside of work. Show appreciation and gratitude. These small touches can make a big difference. 

Learn more: AMP CEO Alexis George on authentic and adaptable leadership

5. Use your voice to build trust: For high-stakes conversations, giving feedback, addressing tension, and leading change, don’t default to AI-generated communications. As AI becomes more widely used, readers look for the personal touches to see if you took the time to write back or if you handed the task over to AI. Personal presence builds trust in ways AI can’t. Employees want to be treated with respect and importance, and AI-generated messages tell them they aren’t worth your time.  

Don't outsource what matters

AI can write your emails. It can restructure your documents, reword your updates, and even mimic your tone. But communication isn’t just about output. It’s about impact. 

In a working world where more than half of our day is spent communicating, the real competitive edge lies not in automation, but in intentionality. 

Your team needs more than polished prose; they need clarity, context, and connection. They need to know you mean what you say and have considered how it might be received. 

So, use AI to save time. But don’t outsource the parts that matter. In the end, what makes professional communication effective isn’t what you send. It’s what people receive. 

And AI can’t read the room. You still can. 

Dr Kelsey Burton is a Lecturer in the School of Management and Governance and has more than ten years of experience in leadership consulting and coaching in the United States and Australia. Dr Gabi Nudelman is a Senior Lecturer in Business Communication in the School of Management and Governance, and she previously worked at the University of Cape Town, teaching business and academic communication across the Faculties of Engineering and the Built Environment, Commerce and the Graduate School of Business.