5 Mistakes Managers Make With AI – And What to Do Instead

The role of AI in management is getting a lot of attention, and rightly so.  Used well, AI can help managers prepare better, communicate more clearly and think through work faster.  Used badly, it can flatten feedback, weaken judgement and make leadership feel inauthentic.  

This matters not just because managers are using AI themselves, but because manager behaviour has a huge influence on whether AI gets used well within a team.  Gallup found that employees who strongly agree that their manager supports AI use are twice as likely to use AI frequently, and nearly nine times as likely to strongly agree that it helps them do what they do best every day.

We explore this in more depth in our AI for Managers workshop, where the focus is not on automating management, but on helping managers use AI well without becoming generic, avoidant or less human.

In this blog, we discuss:

• Best Practice for Using AI in Management
• The 5 Biggest Mistakes Managers Make With AI
• What to Do Instead

Best Practice for Using AI in Management

There’s no doubt that there are benefits to using AI in management. Managers can use AI well to prepare for difficult conversations, sharpen feedback, structure team communication and pressure-test plans. Microsoft’s 2025 Work Trend Index found that people also turn to AI for things humans cannot provide, like 24/7 availability, speed and idea generation.

AI can even be used to support managers’ learning and development. For example, we created a ‘Pocket Coach’ AI assistant for participants on one of our ongoing Manager Development programmes.  It literally gives them a ‘coach in their pocket’ for whenever they need support.  We trained it on the content of the programme, so managers were able to access tailored AI support – even on nightshifts when there was nobody else around to act as a sounding board for them in a tricky situation.

However, there are limits to how much we should be relying on AI in management. For example, managers should never use AI to avoid difficult conversations, fake empathy, replace judgement or make sensitive people decisions for them. AI can support managers, but it should never replace the human qualities that make a great leader.

Best practice for AI in management means striking the balance between technological efficiency and personal judgement – and never losing the human touch.

The 5 Biggest Mistakes Managers Make With AI

The most common mistakes people make when using AI in management are:

• Using it to avoid difficult conversations
• Turning feedback into bland corporate language
• Using it to fake empathy
• Trusting it more than context
• Mistaking speed for quality

Why is Using AI to Avoid Difficult Conversations a Mistake?

AI can help managers prepare for a difficult conversation, but it cannot replace the judgement, clarity or courage really needed.  I’m always amazed at the lengths people will go to in order to avoid a tricky conversation – and for some managers, AI might feel like the easy way out.

One of the biggest traps of AI in management is using it to dodge the uncomfortable parts of being a people leader.  Picture this: a manager has concerns about performance, behaviour or tension in the team, so they ask AI to draft a message, soften the feedback or tell them what to say.  That may seem efficient but in reality, it means the manager is avoiding the thinking that this conversation actually requires.

What to do instead:

• Use AI to clarify your key message
• Ask it to spot vague wording or unnecessary waffle
• Use it to anticipate likely reactions or questions
• Then decide what matters and have the conversation in your own words

Used well, AI can help you organise your thoughts before a tough conversation. But it can’t do the emotional intelligence bit for you – noticing your own reaction, staying steady under pressure and delivering difficult feedback in a way that is clear without becoming clumsy or cold. We talk more about that in our piece on emotional intelligence in the workplace.

Why does AI sometimes make manager feedback worse?

AI is very good at making feedback sound smoother, but not necessarily clearer or more useful.

A lot of managers start with a real point and then ask AI to “make it professional” or “make it sound better”. Seconds later, they have a paragraph of beautifully polished nothing.  Feedback only works when it is clear, specific and rooted in observable behaviour.  If it could apply to almost anyone, it probably will not help anyone.

What to do instead:

• Start with the actual example, not the polished version
• Note what happened, what impact it had and what needs to change
• Use AI to tighten wording or check tone
• Keep the substance rooted in real behaviour you have seen

Radical candour requires bravery and a willingness to give challenging feedback because we care.  Outsourcing this to AI, well…  it doesn’t demonstrate care in the right way.

Why is using AI to fake empathy risky?

AI can imitate the language of empathy very easily, but it cannot genuinely understand what support is real, what the person needs or whether your words match your actions.

Managers are under pressure to sound supportive and composed, especially in awkward or emotionally loaded situations.  So the temptation is obvious: ask AI to make a message sound warmer, kinder or more human.  The problem is that people can usually tell when something sounds polished but emotionally vacant.  That gap is where trust starts to wobble.

What to do instead:

• Think first about what the person may be worried about
• Be clear about what support is accessible
• Use AI to make wording clearer, not to manufacture sincerity
• Only send language you can genuinely stand behind

Gallup’s findings, referenced earlier in this blog, matter here too. Employees are much more likely to use AI meaningfully when manager support feels real and relevant, not vague or performative.

If you want a quick, practical sense-check on that line, our resource on how to use AI as a manager: 9 red flags to avoid is a useful companion.  It highlights the warning signs that your use of AI in management is starting to blur accountability, dilute your voice or weaken trust.

Why shouldn’t managers trust AI more than context?

AI only knows the words and context it has been given.  It does not know your team history, the personalities involved or the political landmines buried under what looks like a simple issue.

AI can be a brilliant thinking partner.  It can help compare options, spot patterns and organise information.  But managers get into trouble when they treat it as if it has a view on the situation, rather than a view on the text in front of it.

What to do instead:

• Keep accountability for people decisions firmly with the manager
• Treat AI as one perspective, not the authority
• Use it to challenge your first instinct or surface missing questions
• Sense-check outputs against the person, the history and the wider context

Why is mistaking speed for quality such a common AI trap?

AI often saves time, and it is very easy to confuse “that was quicker” with “that was better”.

AI can draft an update in two minutes, summarise notes, suggest a response and organise your thoughts.  Useful? Yes. But automatically better? Not even slightly.  In management, speed without thought can create bland communication, sloppy judgement and a creeping sense that the manager is present in meetings but not especially present in the work.

What to do instead:

• Spend the saved time on better questions, clearer thinking and stronger follow-through
• Use the time AI saves to improve quality, not just speed
• Personalise the message before you send it
• Review summaries and recommendations properly

How Should Managers Use AI Instead?

AI in management should be used to support the thinking behind good leadership, not to replace the human responsibility of managing people.

That means using AI to:

• Prepare for difficult conversations
• Sharpen feedback
• Structure team updates
• Organise messy thinking
• Pressure-test plans and priorities

It does not mean outsourcing courage, empathy, judgement or accountability.

Final Thoughts on AI in Management

The most useful question relating to AI in management is not “what can AI do for me?” but “what should I still be doing myself?”

Used well, AI can help managers prepare better, think more clearly and communicate more intentionally.  Used badly, it can make management more generic, more avoidant and less human.  And because managers play such a big role in whether teams see AI as useful, relevant and safe, the impact goes beyond one manager using it badly.

The goal is not faster management. It is better-supported management.

Frequently Asked Questions About AI in Management

If you’re just starting to use AI in management and are still figuring out the best ways to do it, here are a few of the questions that tend to come up most often.