This will sound familiar to you.
You receive an email asking you to complete a task by a certain date. Let’s say the task is updating a specific file. The sender, maybe your boss, tells you it should be quick and easy. There is little explanation of what the file contains or what it will be used for.
You open the file, you think you understand it, and you update it. When you review it together, you realize you misunderstood the task. The work was wrong, and you have to start again from the beginning.
The Cost of Missing Context
A few months ago, I was the manager who failed to provide enough context. I asked my team to update a file without explaining the bigger picture. I had assumed everyone knew what we were talking about and what the information would be used for. They did not. We had to do the work twice and lost valuable time, not to mention the frustration.
Is AI Making Us Worse Communicators?
Many people are raising concerns about the impact of Artificial Intelligence (AI) on our communication skills. I understand this view. We now have tools that let anyone polish their written language. As a result, emails and blog posts are starting to sound the same: “GPT-ish”, as a colleague of mine puts it. People pay even less attention to how they write, because AI will make it look better. This is especially true when they write in a foreign language.
I agree this is a risk. But what if AI can be a catalyst to make us better communicators in spite of that? Not because we learn from how the machine writes (which I do not think everyone is doing), but because of what it requires from us to get the most out of the AI tools.
Let me explain it.
Prompting as a Form of Communication
What is the most important factor in getting good results from Large Language Models (LLMs) today? Prompting.
We have all lived this. You ask CoPilot to write an email to a supplier about a late delivery, with no further detail. It returns something generic and overly polite, in a tone that is not yours, and it misses what actually matters.
So you correct it.
You explain the history with this supplier, you specify how firm the tone should be, you clarify what you need from them and by when. By the third or fourth iteration, you finally get something useful. Total time invested: double what it would have taken to write it yourself.
Prompting is a form of communication too.
We may be getting lazier about finding the perfect, poetic word. However, we are spending much more time explaining exactly what outcome we expect and what context must be taken into account to get the job done right.
Interestingly, the machine is teaching us what we never learned to ask of people. Prompting has become almost a science, and that science boils down to one rule: explain clearly what you want. This is not only a requirement of AI. People need exactly the same. There are three things we demand from AI that we often skip with our teams: context, expectations, and constraints.
Effectiveness Over Elegance
Speaking beautifully or correctly is not the same as communicating effectively. In business, effective communication means clear instructions that leave no room for assumption:
- Provide context. Explain the “why” behind the task. Instead of “Please fill in this Excel“, write “I need this Excel by Friday for a meeting with my boss. I want to show that we have grown month on month.” With that, your team knows which data matters and which they can ignore. Without it, they make reasonable decisions that turn out to be wrong, and they do not even know it.
- Set expectations. Define what the final output should look like. Instead of “Can you summarise the report?“, write “I need three key points, two slides maximum, in language someone outside the industry can follow.” Without this, everyone imagines a different result. When five different versions of the same task come back, the problem was not the people. It was the instruction.
- Define constraints. Make clear what should not be done. Instead of “Please launch an RFQ for this component“, add “Invite only suppliers already qualified for this part, do not share our target price, and do not mention volumes beyond next year, because the forecast is not approved yet.” Without these limits, a buyer can run a perfectly organised RFQ and still weaken our negotiating position or create expectations we cannot meet.
So here is the question I keep asking myself: why do we give more explanation to an algorithm than to someone who has worked with us for years?
With AI, we provide context, define the format, and set clear limits, because we know it will fail otherwise. With people, we send two lines and assume they already know. Then we are surprised when something goes wrong.
The precise word is valuable. But context, expectations, and constraints are what actually change results. We have learned to apply them on the screen. Now we need to apply them off it.
Tomorrow morning, before you send that first email asking your team for something, pause for a moment. Remember that you are addressing the most powerful machine in the world: another person’s mind. Give it the context it deserves.