What Executive Coaching Gets Wrong Without Practice Built In

Executive coaching works through conversation. A coach and a leader talk through a situation, a decision, a pattern of behavior, and the leader leaves with better judgment about what to do. That model has held up for decades because it addresses something real. Executive coaching and practice rarely sit in the same program, though, and that separation is where the model breaks.

The coaching hour stops at the edge of the conversation. The leader leaves knowing what to say. The next time they are in the room where it matters, they still have to say it, out loud, to a person who reacts, and nothing in the coaching hour rehearsed that part.

The Gap Between Deciding What to Say and Being Able to Say It

Coaching produces a plan. “Open by naming the behavior specifically. Do not lead with the org change. Give them room to respond before you move to next steps.”

Then the leader walks into a performance conversation with someone who has worked for them for four years, whose face changes at the second sentence, and the plan dissolves. They soften the opening. They bury the specific behavior inside a compliment. They talk longer than they meant to because the silence is unbearable. They leave the room having delivered something adjacent to what they intended, and the direct report walks out unsure whether anything was actually said.

The leader knew what to do, so judgment held. The delivery did not hold under load, and delivery under load is a motor skill more than an intellectual one.

Ericsson’s original work on expertise named the missing ingredient: effortful, repeated practice on a specific sub skill with immediate feedback. Macnamara and Maitra revisited the framework and re examined how much of expert performance deliberate practice accounts for, and the debate about magnitude has not displaced the core mechanism. Repetition on the specific sub skill, with feedback, is how a behavior becomes available under pressure. A coaching conversation gives the leader reflection, and reflection alone leaves the sub skill untrained.

The Leaders Who Need It Most Are the Ones Coaching Never Reaches

Executive coaching is expensive and allocated at the top. In most organizations that means a small population of senior leaders get a coach, and everyone below them gets a workshop.

The conversations do not respect that allocation. A first time manager delivers their first performance warning within months of being promoted. A director announces a reorganization they had no part in designing. A frontline supervisor has to tell someone their role has changed. A mid level leader presents to an executive team that is skeptical before they walk in. These are the highest stakes conversations in the company by volume, and they are being run by the people with the least preparation.

Those leaders typically get one of two things. A workshop, where they discuss the conversation and possibly watch a demonstration. Or nothing, and they learn by getting it wrong with a real person who remembers it. Ochsner Health approached this directly by working on frontline leadership skills across critical conversations, which is the right population to target. Frontline leaders touch more employees than the executive team does.

VML built leadership development on AI roleplays and won Brandon Hall awards for the program, which matters mostly as evidence that practice based leadership development is a real program design and holds up as one when judged against peers.

What Practice Actually Looks Like for a Leadership Conversation

The instinct is to practice the whole conversation end to end. That is the least efficient version, because the part that breaks is usually under a minute long and sits in a predictable place.

Drill the moments where leaders lose the thread:

• The opening sentence of a hard conversation, where vagueness gets introduced and never recovered.
• The first emotional response from the other person, which is where most leaders start negotiating with themselves.
• The moment someone asks a question the leader cannot answer, usually about the decision behind the decision.
• The silence after the leader has said the difficult thing, which most people fill with qualifiers that undo it.
• The close, where a conversation that went well produces no agreement about what happens next.

Each of those is a short drill that can be run many times. Run the opening sentence of a performance conversation ten times against a persona that reacts differently each time, and the leader develops something they can produce under pressure. That is a different exercise than discussing the opening sentence in a workshop.

Feedback a leader can act on

The feedback has to be specific to be worth anything. “You came across well” gives the leader nothing to change. “You named the behavior in your fourth sentence instead of your first, and you used the word maybe three times” gives them two things to fix on the next run. Structured feedback models help here, and the SBI feedback model gives leaders a frame for both receiving that feedback and delivering it themselves.

Where AI roleplays sit next to the coach

This is where AI roleplays for leadership development fit, as a layer that sits next to coaching. A coach cannot be available for the twelfth repetition of a ninety second opening, and would be poorly used if they were. The coach’s time belongs on judgment, on the situations that are ambiguous, on the pattern the leader cannot see in themselves. Repetition belongs somewhere it can happen at two in the morning before the conversation that is scheduled for nine.

Yoodli covers that repetition side. A leader picks the conversation, runs the opening against an AI persona that pushes back, and gets feedback on the specifics: whether the behavior got named, how many hedges crept in, whether the pause held. Then they run it again.

Where This Fits in a Leadership Program You Already Run

Most leadership development programs already have the content. Frameworks for feedback, for difficult conversations, for change communication. The content is rarely the weak part.

Put practice between the content and the real conversation. A leader completes the module on performance conversations, then runs the conversation against a persona, gets scored, runs it again, and only then has it for real. The sequence matters more than any individual piece. Content without practice produces leaders who can describe a good performance conversation. Skip the content and you get confident delivery of the wrong approach.

For programs built around crucial conversations, the practice layer is what converts a well attended workshop into behavior change that shows up in an engagement survey six months later.

The Objections You Will Hear

Three push back reliably when you propose adding practice to a leadership program.

The first is that leadership conversations are too situational to rehearse. Every performance conversation is different, so practicing one teaches you nothing about the next. The response is that the rehearsal targets a sub skill: naming a behavior specifically, holding a pause, responding to a defensive reaction without retreating. Those transfer across situations because they are the parts that are the same every time.

The second is that senior leaders will not do it. Some will not, and that is worth accepting early. The population where practice has the most effect is first time and frontline managers, who have the most conversations, the least preparation, and considerably less ego invested in the idea that they already know how.

The third is that AI roleplay practice will feel artificial and produce artificial behavior. It does feel artificial for the first attempt or two. So does every rehearsal. The question to test is whether a leader who has run the opening ten times delivers a better opening for real, which is measurable inside a single cohort.

Can Executive Coaching and Practice Work in the Same Program?

Yes, and the two get better when they are scheduled together. The coach owns judgment: which conversation to have, what the leader keeps avoiding, why the same pattern shows up in three different relationships. The practice layer owns repetition: the opening sentence, the pause, the response to pushback, run until it holds.

The simplest way to connect them is to make practice the homework between coaching sessions. The coach and leader agree on one conversation to work on. The leader runs it as an AI roleplay several times before the next session and brings the scored attempts with them. The coaching hour then starts from what the leader actually said under pressure instead of what they remember intending to say. Coaches tend to like this arrangement once they see it, because it gives them evidence to work from and moves the drilling out of their calendar.

For leaders below the coaching tier, the same sequence runs without the coach. A manager, a program module, and a practice assignment between the module and the real conversation.

What to Measure

Leadership development gets measured on satisfaction scores, which is why it struggles to defend a budget. Better options already exist in your data.

Track practice completion and score improvement across attempts, which tells you whether the program is doing anything at all. Then connect it to numbers your HR and people teams already report:

• Voluntary regretted attrition on teams whose managers completed the program against those who have not.
• Manager effectiveness scores in your engagement survey.
• Time to productivity for new managers.
• Internal promotion rate into leadership roles.
• Employee relations case volume, which often moves when frontline managers get better at conversations that were previously escalated.

At the conversation level, score whether the leader named the issue specifically, whether they gave the other person room to respond, and whether the conversation ended with a stated next step. Those three are observable and they are the ones that fail most often.

If you are running executive coaching for your senior leaders, keep it. Add a practice layer for everyone below them, starting with the two conversations your managers escalate most often. The leadership page shows how that layer is usually structured.