Use AI to save time, then spend that time being human.

I was recently BCC'ed into an email sent to another facilitator, asking them to provide a quote.... Think about this from your persective... You get an email from a potential new client, but its not actually addressed to you. It's addressed to your competitor, and you are secretly attached.

It was an odd position to find myself in. I could see the request and the conversation, but the other coach could not see that I was there. I was not entirely sure whether I was meant to respond, observe or simply understand that another option was being explored.

When I raised it, the explanation was that it had been done for efficiency.

Rather than having separate conversations with different people, one email had been sent, with everyone involved in some way. From the sender’s perspective, it was the simplest and fastest way to gather the information they needed.

And I do understand that.

We are all busy. We are managing too many emails, too many decisions and too many competing priorities. Sending one message instead of several can feel like an obvious way to save time.

But it made me think about what we actually mean when we describe something as efficient.

Because while sending one email may have been efficient for the sender, it created confusion for the people receiving it. It blurred the boundaries of the conversation, made the intentions less clear and ultimately required further communication to understand what was happening.

It saved time in one place but created more work somewhere else.

Which raises an important question:

Who gets to decide what efficiency is?

We often treat efficiency as though it is an objective measure. Fewer words. Fewer emails. Fewer conversations. Fewer minutes spent completing a task.

Job done. Box ticked. Move on.

But efficiency is not always about how quickly information leaves our hands. Sometimes, it is about how easily it can be received, understood and acted upon by another human being.

An action that saves me thirty seconds but creates ten minutes of confusion for someone else is not necessarily efficient.

A process that reduces the number of conversations but leaves people feeling disconnected, dismissed or unclear may not be efficient at all.

It may simply be fast. And fast and efficient are not always the same thing.

The computer’s definition of efficiency

As artificial intelligence becomes more embedded in our lives and workplaces, this distinction matters more.

Computers are extraordinary at efficiency. They can process information, identify patterns, organise ideas, summarise documents and generate responses in seconds. They remove repetition, streamline administration and help us begin tasks that might otherwise sit untouched.

I use AI. It has the potential to make our work significantly better.

But a computer’s definition of efficiency is usually based on speed, accuracy, volume and task completion.

Humans are more complicated.

We interpret tone, intention and context. We notice what has been said and what has been left unsaid. We consider our history with the person and what their communication might mean for the relationship.

A concise message can still create confusion. An accurate response can still feel cold. A task can be completed while the human need underneath it remains unmet.

ChatGPT does not need trust, reassurance, dignity or connection.

People do.

Not every interaction is a transaction

One risk of a more automated world is that we begin to treat every human interaction as a transaction.

Question asked. Answer supplied.

Problem identified. Solution generated.

But conversations at work are not only about exchanging information. They are how we build relationships, understand context, uncover concerns and repair misunderstandings.

Sometimes the conversation is not an obstacle to the outcome.

Sometimes the conversation is the outcome.

• A phone call may reveal uncertainty that an email never could.

• A leader taking time to explain a decision may communicate respect.
  

• A conversation with a potential client or collaborator may help both people decide whether there is enough trust and alignment to work together.

That cannot always be captured in a form, automated through a workflow or compressed into three lines.

Efficiency for whom?

Before we call something efficient, we should ask:

Efficient for whom?

• Is it efficient for the sender but confusing for the receiver?

• Is it efficient for the organisation but frustrating for the customer?

• Does it remove unnecessary work, or simply move the work onto someone else?

There are processes that need to be faster. There are meetings that could have been emails and reports nobody needs.

Efficiency matters.

But we should not remove the human moments that make the work possible: the informal check-in before a difficult conversation, the explanation behind a decision, the opportunity to ask a question or the coffee that builds trust for future work.

These moments may look inefficient on paper, but they can prevent hours of future conflict and confusion.

More importantly, they remind people that they matter.

AI should remove friction, not humanity

AI should not help us avoid people. It should help us spend more meaningful time with them.

• It can draft the agenda, but we still need to notice who has not spoken.

• It can summarise feedback, but we still need to listen to what people are telling us.

• It can write the first email, but we still need to consider how it will be received.

• It can help leaders prepare for conversations, but it cannot take responsibility for having them.

The best use of technology is not to make every interaction shorter. It is to remove low-value work so we have more capacity for the conversations that matter.

Real humans may feel inconvenient compared with instant answers and automated summaries. They misunderstand, change their minds and bring emotion into logical decisions.

They need context, reassurance and sometimes a second conversation.

That is not inefficiency.

That is humanity.

A few things to try

Before you send, ask yourself who the communication is efficient for.

1. Consider whether a short conversation would create more clarity than a fast email.
   

2. Use AI to reduce administration, not to remove judgement, empathy or connection.

   
   

3. Check whether your message gives people enough context to understand what is happening and what is expected of them.

   
   

4. Remember that not every interaction is a transaction. Sometimes the conversation is the work.

   
   

5. Pause before choosing the fastest option and ask whether it is also the clearest, kindest and most effective one.

Humans are not ChatGPT.

We are messy, emotional and relational. We need space to explain ourselves, ask questions, repair misunderstandings and change our minds.

That may not always be the fastest way to work.

But sometimes, it is the most efficient way to be human.

Blue Mercury Leadership helps people become more human at work.

Through practical leadership development, coaching and tailored team programmes, we help leaders build stronger relationships, have better conversations and create cultures where people can do their best work.