I want to hear from you

I am getting frustrated with a certain kind of conversation. I ask someone a question and receive a long, polished reply. It sounds professional. It covers several angles. Yet I finish reading it without knowing what that person actually thinks.

Sometimes it feels as though I am communicating with an agent, with a person somewhere behind it pressing send. There are more words, but less of the person I wanted to speak to.

I understand the appeal. AI can turn a rough thought into a clear message in seconds. It can help when we are busy, tired or writing in a language that is not our first. Those are useful things. My frustration starts when the generated message becomes the whole contribution.

A polished answer can still miss the point

Imagine asking a colleague which option they would choose. The reply explains both options, lists the advantages and ends with a careful conclusion that the decision depends on your priorities. That may be a reasonable first draft. It still leaves you waiting for your colleague’s judgement.

Or imagine raising a concern and receiving a message full of reassuring language. It sounds considerate, but it does not address what worried you. Someone has produced the shape of a response without doing the work the conversation needed.

The result does not have to be obviously wrong to be frustrating. It can be accurate in parts, fluent throughout and still miss the question. Then the recipient has to ask again, clarify the context or work out what the sender meant. Saving time on one side can create more work on the other.

I do not need every message to be perfectly written. I need it to be useful, honest and connected to the person sending it. A short answer with a clear opinion often gives me more than several paragraphs that could have come from anyone.

We can make ourselves disappear

We spend a lot of time worrying that AI will replace us. Yet there is a smaller, everyday version of that problem which we create ourselves.

If we hand over every email, every message and every response, then send the result without questioning it, what are we contributing? We become the delivery mechanism for the machine’s output. We keep our name on the message while removing our thinking from it.

That is what I mean when I say we start becoming agents ourselves. We pass requests in one direction and generated answers in the other. Our opinion, experience and understanding of the person on the other side become harder to see.

This is the part that troubles me. We are afraid of being eliminated by AI, while making choices that remove us from our own work and relationships. We should pay attention to how often we do that.

Our judgement belongs in the work

Strategic thinking and emotional intelligence are strengths we should bring to AI-assisted work. We can understand why a question matters, what a relationship needs and which decision we are prepared to stand behind.

An AI system can suggest a strategy or produce language that sounds empathetic. We still have to judge whether the strategy makes sense and whether the words fit the situation. The message goes out under our name. We are responsible for what it says.

Sometimes the right response is a direct answer. Sometimes it is an honest admission that we do not know. Sometimes it is a difficult conversation that should not be hidden inside a beautifully worded email.

That judgement needs to be present from the beginning. Before asking AI to write, we should know what we want to say, why we want to say it and what the other person needs from us.

Make the result yours

A useful prompt begins with our thinking. Give the system the actual question, the relevant context and the position you have reached. If you are still deciding, ask it to challenge your reasoning or explore options. Use that exchange to reach your own view.

Then read the output critically. Does it answer the question? Are the facts right? Does it make a promise you cannot keep? Has it added a confident sentence that you would struggle to explain?

And make corrections. Remove the paragraph that says nothing. Change the sentence that sounds nothing like you. Add the detail that matters to this particular person. If the answer is wrong, rewrite it.

Corrections are part of the work. They are also how we learn. The more clearly we can explain what is wrong with a draft, the better we become at setting the brief and judging the next result.

Build that skill instead of getting used to accepting the first version. Being able to produce a message quickly is useful. Being able to recognise and improve a poor message is useful too.

Use AI to become more yourself

I want AI to help us be more creative and productive. It can offer a starting point, question an assumption, help organise a complicated thought or give us room to explore an idea we would otherwise leave unfinished.

We can use that help while keeping our own voice and judgement in the result. The time we save can go into understanding the problem, making a better decision or having a conversation that deserves our attention.

I want to see more of the person behind the work: their ideas, their decisions, their ability to notice when something is wrong. When I receive a message from someone, I want to feel that they have been present in writing it.

We are not agents. Let’s use AI to boost our creativity and productivity, and take responsibility for what we send. Before pressing send, read the message and ask: does this say what I think, and am I prepared to stand behind it?