What Is Coaching Leadership? Definition and Examples

How easy is coaching?

First and foremost, most leaders feel they are great at it. Though, most are not!

It is a skill that is never complete and you will be learning all the time.

Key Takeaway: Coaching is harder than most leaders assume, and coaching skill is never complete because a leader keeps learning all the time.

How can we be more coach-like?

An approach that worked for me is rather than advising all the time I now ask some of these questions in at least one of my daily conversations. Why not give this a try?

• What is on your mind?
• And what else?
• What is the real challenge here for you?
• What do you want?
• How can I help?
• If you're saying yes to this, what are you saying no to?
• What was most useful for you?

The seven questions above come from The Coaching Habit by Michael Bungay Stanier. The one most leaders drop first is "And what else?". Ask it two or three times before you offer an opinion, because the first answer a person gives is usually the tidy one and the useful answer arrives after it.

Key Takeaway: A leader becomes more coach-like by replacing advice with the seven questions from The Coaching Habit and by asking "And what else?" two or three times before offering an opinion.

Coaching vs mentoring

In some cases, I have found that I still need to mentor. For example, if someone has the skill already then they can be coached by asking the right questions and they can be inspired to take action.

For example, if someone has expertise in accessibility testing, a leader can use coaching to use their previous experience to make a difference for the team.

In cases where they do not have the skills, showing them how to act on their request can help. Mentoring helps in this scenario, for example mentoring someone on how to write and run scripts in Jmeter when running performance testing scripts.

I will coach when:

• The person has the ability.
• They have worked in a similar situation before.
• They will get more value from seeing the task from their own perspective as it needs deep thinking.

I will mentor when:

• There is an ability gap
• It is the first time they are trying a new task
• It is a complex task where working and pairing add more value.

I may also mentor and coach at the same time, they are not mutually exclusive. From my experience sometimes doing both at the same time can help to teach a tricky skill that is hard to memorize.

The coach or mentor decision has become harder now that engineers work alongside AI assistants. A person can produce working code, tests, or scripts on request, so finished output no longer proves the skill is there. Ask the person to walk you through what the generated code does and what it does not cover before you choose. If they can explain the risk it misses, coach them. If they cannot, mentor them, because the ability gap is real even though the task looks complete.

Key Takeaway: Coaching suits a person who already has the ability, mentoring suits an ability gap or a first attempt at a task, and a leader can use coaching and mentoring together on one tricky skill.

When does coaching leadership not work?

Coaching leadership fails in three situations: the person does not want to be coached, the leader has no real expertise in the work, or the moment needs a decision rather than a conversation. Daniel Goleman set out the first two limits in Leadership That Gets Results, the 2000 Harvard Business Review article that named coaching as one of six leadership styles. His research found that leaders used the coaching style least often of the six, and the reason they gave was that they had no time for the slow work of developing people. Goleman also wrote that coaching makes little sense when employees are resistant to learning or changing their ways, and that a second risk appears when the leader lacks the expertise to help the employee grow.

Two of those situations turn up often in engineering teams. A live production incident is the clearest one. During an outage you direct, you do not ask "And what else?". Run the coaching conversation afterwards in the review, when the pressure is off. The second is the culture around you. If the organization still rewards leaders for holding the answer, coach-like questions get read as indecision, and the team learns to wait for the instruction anyway.

The time constraint Goleman described is also why AI coaching tools are now built and sold to leaders. The International Coaching Federation publishes an AI Coaching Framework and Standards for the providers and organizations adopting them, and it positions AI-driven coaching as a complement to human coaching rather than a replacement for it. Use a tool to prepare your questions or to write up your notes after a session. Hold the conversation yourself.

Key Takeaway: Coaching leadership does not work when the person resists being coached, when the leader lacks expertise in the work, or when the situation demands an immediate decision such as a live production incident.

How does AI change coaching leadership?

AI changes two parts of the job: how a leader prepares for a conversation, and how a leader judges whether someone actually holds a skill. It does not change who is accountable for the conversation.

On preparation, AI coaching products now sit between a leader and a session, and the International Coaching Federation has written standards for them. The Artificial Intelligence Coaching Framework and Standards sets out six domains: foundational ethics, co-creating the relationship, effective communication, learning and growth facilitation, assurance and testing, and technical factors such as privacy and accessibility. ICF also publishes an AI Coaching Standards self-scoring worksheet so a leader or a buyer can rate a tool against those domains before adopting it. Read the stated limits first. The standards are written as principles rather than rules, they name bias and confidentiality as the ethical risks, and they treat AI as a complement to human coaching rather than a replacement for it.

On judging skill, the shift is sharper for engineering leaders. An engineer working with an AI assistant can hand you a passing test suite without holding the knowledge behind it, so completed work is no longer evidence of ability. The check is the one you would use in a code review. Ask what the output does not cover, and listen for whether the person can name the gap.

Use these tools for the parts of coaching that were always administrative, such as scheduling, drafting questions before a session, and writing up notes after it. Keep the conversation itself. A tool can prompt you to ask "And what else?", but it cannot carry the accountability that makes the answer worth giving.

Key Takeaway: AI helps a coaching leader prepare and take notes, but ICF standards treat AI coaching as a complement to human coaching, and finished AI-assisted work no longer proves an engineer holds the skill.

Closing

We should all try to be more coach-like. For me, reflecting, reviewing, and adjusting my approach has been key, and I have learned by not always advising and dictating.

We should do this as it means that team members will grow and they will be happier.

It increases diversity of opinions, conversations are better, and it is a win-win all round.

Coaching is an extra skill you can add to your toolkit as a leader, so give it a try!