Daniel Costa manages an eight-person customer support team at a logistics software company. For two years he told himself the same thing every manager tells themselves: "I'll get to development planning once things calm down." Things never calmed down. Then one of his strongest agents, Priya, handed in her notice. In the exit conversation she said something that stuck with him: "I never knew where I was going here. Nobody ever talked to me about growing." Daniel realized he had been so busy keeping the queue clear that he had stopped developing the very people who kept it clear. This lesson follows how he rebuilt his approach to coaching and development planning, using AI to do the heavy lifting on structure while keeping the human part, the belief and the relationship, firmly his own.

What this lesson covers

Development planning is the work of helping someone become more capable over time, not just perform their current job well. This lesson shows you how to use AI as a thinking partner to spot development needs, design realistic learning plans, and prepare coaching conversations, while you stay the actual coach. The AI can synthesize input and draft a plan in minutes. It cannot believe in your team member, build trust with them, or judge whether the timing is right. That part stays yours.

We will work through the difference between performance and development, the five things that actually make people grow, the different modes of coaching and when to use each, and how to build a development plan that holds up. Daniel's story carries us through, and we close with a full worked plan for one of his team members using two well-known frameworks: GROW for the coaching conversation and 70-20-10 for the learning design.

Why good managers skip this work

Almost nobody decides not to develop their team. They just never get to it, and the reasons sound perfectly sensible from the inside. "I do not have time" is the most common, and it is true right up until you lose someone. "They have to own their development" is the second, and it contains a half-truth that makes it dangerous: people do have to own their growth, but they cannot own something nobody has ever shown them. "I do not know how to coach" is the third, and it is the most honest, because coaching is a learnable skill that most managers were never actually taught.

Set against those three reasons is one fact that outweighs all of them. Developing your people is among the most important things you do, and it is what people remember. Nobody leaves an exit interview talking about the ticket queue. They talk about whether anyone helped them become more than they were when they arrived. That was Priya's whole message, and it is why Daniel treated the excuses as excuses rather than constraints.

Development is not the same as performance

The first thing Daniel had to untangle was a confusion that had quietly stalled his whole team. He had been treating his quarterly reviews as development conversations, but they were really performance conversations wearing a development costume. Performance is about the job someone does today and how well they do it. Development is about who they could become and what they need to learn to get there. They are related, but they are not the same.

This distinction matters because the two can move in opposite directions. Daniel had a steady agent, Marcus, who hit every metric and had zero interest in changing anything about his role - a strong performer with no development appetite. He also had a newer hire, Ana, who fumbled some tickets but had obvious raw potential and was hungry to grow - a developing person with current performance gaps. If Daniel had only looked at performance numbers, he would have invested in the wrong person. Development planning forced him to look at trajectory, not just the current scoreboard.

What actually makes people grow

Before designing any plan, Daniel needed to understand the conditions under which people actually develop. People grow when five things are present at once. They see why it matters, meaning they have a clear picture of the next level and want it. They get stretch, because nobody develops inside their comfort zone. They get honest feedback, so they understand where they really stand. They get support, which means a coach, resources, and protected time. And they have agency, because growth that is forced on someone rarely sticks.

Daniel found it useful to treat these five as a checklist. When development stalls, usually one of them is missing. Ana had stretch and feedback but no clear vision of where she was heading, so her effort felt aimless. Marcus had every condition available to him except genuine interest, which is why pushing him would have been a waste of everyone's energy. You cannot want someone's growth more than they do.

Choosing how to coach in the moment

Coaching is not one fixed style. Daniel learned to shift between four modes depending on what the person in front of him needed. Directive coaching, where you simply tell someone what to do, is useful early on when someone genuinely does not yet know the ropes. Facilitative coaching, where you ask "what do you think you should do?", builds the person's own judgment and agency. Developmental coaching keeps the focus on long-term growth rather than the immediate task. Supportive coaching is mostly about steadying someone emotionally when they are struggling.

The mistake Daniel used to make was running every conversation in directive mode because it was fastest. With Ana, who was capable but unsure of herself, defaulting to "here is what you do" kept her dependent on him. Switching to facilitative questions - "what options do you see here?" - was slower in the moment but built the independent thinking she actually needed.

It is worth separating two words that get used interchangeably. Coaching is helping someone think through their own development; the answers are theirs and your job is to draw them out. Mentoring is sharing wisdom from your own experience; the answers are yours and your job is to make them useful. Both matter, and Daniel used both, but confusing them is how managers end up giving advice when the person needed to be asked a question.

The six situations that call for a plan

Development planning is not one activity, and knowing which situation you are in tells you what the plan should look like. There is high-potential development, where someone has clear runway and you are building toward it. There is skill-gap development, where someone needs to learn something specific to be more effective in the job they already have. There is readiness development, where someone wants a promotion and needs to know precisely what to work on to earn it. There is lateral development, where someone wants to try a different kind of work rather than a bigger version of the same work. There is leadership development, where you are coaching an individual contributor toward managing people, which is a change of job rather than a change of level. And there is recovery coaching, where someone whose performance has dipped needs help getting back to solid ground.

Daniel found that naming the situation first saved him from a common mistake: writing an ambitious high-potential plan for someone who actually needed a narrow skill-gap plan, or treating a recovery situation as though it were a promotion track. The shape of the plan should follow the situation, not your enthusiasm.

Using AI as a thinking partner, not a substitute coach

Here is where AI earns its place. Daniel used to stare at a blank document trying to turn a vague sense of "Ana has potential" into something structured. AI removed that blank-page problem. He could describe the person, their strengths, their gaps, and a rough timeline, and get back a structured first draft of a development plan in minutes.

The boundary he set for himself was firm: AI helps him synthesize and structure, but it never becomes the coach. It can draft a learning plan, surface development angles he had not considered, and help him think through the shape of a tough conversation. It cannot believe in Ana, build a relationship of trust with her, or sense when she is quietly losing confidence. Those are the parts that actually drive growth, and they stay human. As Daniel put it to himself, "the AI can write the plan, but it cannot mean it."

A prompt he reused looked like this: "Help me build a six-month development plan for a support agent moving toward a senior or team-lead track. Strengths: empathy with customers, fast learner. Gaps: handling escalations independently, mentoring newer agents, writing clear internal documentation. Structure it with a vision, current state, learning activities, stretch assignments, the support I provide, milestones with dates, and how we measure progress." The draft that came back was about 80 percent usable. Daniel's job was the other 20 percent: the parts only he knew, like the fact that Ana freezes under public pressure, so any "lead a team meeting" stretch needed scaffolding.

The anatomy of a development plan that holds up

Whether AI drafts it or Daniel writes it himself, a development plan that actually works has seven parts. There is a vision of the growth and what the next level looks like. There is an honest current state and the gap between here and there. There is a learning plan describing what to learn and how. There are stretch opportunities, the real projects or roles that force growth. There is the support the person needs from you. There is a timeline with expected dates. And there are measures, so both of you can tell whether growth actually happened.

Daniel's hard-won rule was about ambition. His instinct was to load the plan with five growth areas because he was excited. That instinct sets people up to fail. He learned to pick one major development focus per quarter, build an early achievable win to create momentum, and remember that behaviors can shift in months while how someone sees themselves shifts over years. Going too fast crushes confidence; going too slow loses momentum. Most managers err toward too fast.

Development is a partnership, and both sides have to show up

The most useful reframe Daniel adopted was that a development plan is a two-sided commitment, and he wrote both sides down. His accountability was to provide the vision of what the next level looks like, to create genuine stretch opportunities rather than just more of the same work, to give honest feedback often enough to be useful, and to supply the support he had promised. Ana's accountability was to actually take on the stretch, to seek and absorb feedback rather than defend against it, to apply what she learned rather than collect it, and to take initiative between their sessions rather than waiting for the next one.

Naming both halves out loud changed the dynamic. It stopped development from being something Daniel did to Ana, which is the version that quietly fails, and it gave each of them something specific to be held to. When a plan stalls, this is the first place to look: one of the two sides has stopped showing up, and it is worth knowing which.

Worked example: Ana's six-month plan with GROW and 70-20-10

Daniel decided to build a real plan for Ana, who wanted to move from support agent toward a senior agent role with eventual team-lead potential. He used two frameworks together. GROW structured his coaching conversation, and 70-20-10 structured how she would actually spend her learning time.

GROW stands for Goal, Reality, Options, and Will (sometimes called Way Forward). It is a four-part structure for a coaching conversation. Daniel walked Ana through it rather than lecturing her:

• Goal. "Where do you want to be in six months?" Ana said she wanted to handle escalations on her own and start mentoring new hires. They wrote it concretely: independently resolve Tier-2 escalations and onboard at least one new agent by month six.
• Reality. "Where are you now, honestly?" Together they noted that Ana currently escalated about 60 percent of Tier-2 tickets to Daniel, and had never formally mentored anyone. That number was the baseline they would measure against.
• Options. "What are the ways you could close that gap?" Rather than dictating, Daniel asked. Ana suggested shadowing a senior agent and keeping a log of escalations. Daniel added the idea of co-handling one new hire's onboarding.
• Will. "What will you commit to, and by when?" Ana committed to specific first steps with dates, which made the plan hers rather than his.

For the learning design, Daniel used 70-20-10, which says roughly 70 percent of development comes from doing challenging real work, 20 percent from learning through other people, and 10 percent from formal training. Spread across Ana's roughly 40-hour week, that works out to about 28 hours of stretch work, 8 hours of learning from others, and 4 hours of formal study. Daniel did not police the hours to the minute; he used the ratio to make sure the plan was weighted toward real practice, not just courses.

• The 70 percent (on-the-job stretch). Ana takes ownership of all Tier-2 escalations for two named enterprise accounts, with Daniel available but not stepping in unless she asks. This is the real challenge that forces growth.
• The 20 percent (learning from others). Ana shadows Daniel's most experienced agent two hours a week and co-runs the onboarding of one new hire, learning mentoring by doing it with support.
• The 10 percent (formal learning). Ana completes a short internal course on de-escalation technique and writing clear support documentation.

The milestones gave the plan teeth. By month two, Ana resolves at least 40 percent of her Tier-2 escalations without handing them to Daniel, up from the starting point where she handled almost none. By month four, that figure reaches 65 percent and she has shadowed her first new hire through their first week. By month six, she independently handles 80 percent of Tier-2 escalations and has fully onboarded one new agent. Daniel's support commitment was concrete and scheduled: a recurring 30-minute coaching conversation every two weeks, on the calendar, not flexible.

Six months later the escalation hand-off rate had moved from 60 percent down to 18 percent, and Ana had mentored two new agents, not one. More telling than the numbers, Ana started raising her hand for things. The plan worked not because the framework was clever but because Daniel showed up to every single one of those biweekly conversations.

When the destination is a leadership role

A year later Daniel faced a different version of the problem. A strong individual contributor on his team was interested in eventually managing people, on a horizon of roughly 18 months. Three gaps stood between them and that role: they did everything themselves rather than delegating, they worked almost entirely inside their own team rather than influencing across teams, and they were excellent at execution but had done little strategic thinking. Daniel described exactly that to an AI assistant and asked how to structure the plan, which gave him a usable skeleton in a couple of minutes.

What he built from it had a distinct shape, because leadership development is a change of job rather than a bigger version of the same one. It opened with a vision conversation rather than a plan: "I see you as a strong candidate for a manager role down the line. Not immediately, because there are things I want to see you grow in. But I think you have it in you. Are you interested?" That question is not a formality. Starting with why matters more here than anywhere else, because plenty of strong contributors pursue management for the title and then discover they wanted the craft. If the interest is not genuine, everything after it is wasted.

The plan then set three focus areas, each with a reason and a checkpoint. Delegation came first, because it is the hardest shift for a strong individual contributor to make: managers multiply themselves through other people, which means deliberately giving away work you could do faster yourself. The checkpoint at month three was a concrete share of work handed off. Cross-team influence came second, because managers have to think beyond their own group, so the stretch was leading a genuine cross-team initiative, with a month-six checkpoint of having shaped a decision outside their own team. Leadership mindset came third and slowest, because it is about how you see yourself and others rather than a skill you can drill. That one ran to month twelve, supported by coaching on real decisions, watching leaders they respected, and mentoring someone themselves.

Two supports made the difference. Daniel committed to a monthly coaching conversation focused only on growth, separate from work status, and he introduced the person to an experienced leader elsewhere in the company as a mentor, because there are things a manager cannot teach about their own job. Peer feedback was built in as well, since influence is measured by how others experience you, not by how you experience yourself. The timeline was explicit and staged: monthly check-ins on progress, a bigger checkpoint at month six to decide whether they were ready for a larger stretch, another at month twelve on management readiness, and a manager role realistically possible around month eighteen.

Coaching someone through a rough patch

Development is not always upward. Three months in, Ana hit a wall after badly mishandling an escalation that reached a frustrated enterprise customer. Her confidence dropped, and Daniel could feel her pulling back. This is where supportive coaching matters more than any plan.

Daniel ran the conversation in a simple shape: he checked in and genuinely listened, he normalized the struggle ("everyone who has ever grown into escalations has blown one - this is the cost of doing the harder work"), he helped her name what was actually hard, he reframed the progress she was discounting, and he recommitted his support. He did not pretend the mistake was fine; he framed it as information, not a verdict. The plan paused for a week, then resumed. Permission to fail is what made the stretch possible in the first place.

The hardest version of this is someone newly promoted into a first leadership role who is overwhelmed and quietly considering whether they made a mistake. The same five moves apply, with two additions. Reduce the pressure by resetting expectations to what is actually achievable in the first months, and teach the thing they do not know, which is usually what management actually consists of, since nobody explained it before handing them the job. Then close by naming a realistic horizon rather than offering vague reassurance: the discomfort is temporary, in a few months it will feel more settled, and within a year the job will feel like theirs. Say it only if you mean it, because people can tell, but say it, because someone in that state genuinely cannot see forward on their own.

The traps that quietly sink development plans

Daniel kept a short list of the failure modes he had personally fallen into, taped inside a notebook. Coaching someone who does not actually want to grow, like Marcus, wastes everyone's energy - always ask first, and accept the answer. Plans that are too ambitious crush confidence - one focus per quarter, build an early win. Avoiding the uncomfortable honest feedback keeps coaching shallow - directness and kindness are not opposites, so be both. Talking about growth without assigning real stretch produces no growth at all - words do not develop anyone, hard work does. And inconsistent coaching, where you commit to biweekly sessions then skip them when you get busy, teaches the person that their development is not actually important. That last one was Daniel's original sin, and fixing it was the single highest-leverage change he made.

A few refinements he added over time. Interest changes, so ask again: what held no appeal for someone eighteen months ago may be exactly what they want now, and a single no is not a permanent answer. Name progress out loud, because people lose confidence when growth is invisible to them, and "you have made real progress on delegation" is not flattery, it is data they cannot see from the inside. Support the stretch rather than just assigning it, which means saying plainly that the work is hard, that failure is possible, and that you have their back, then debriefing afterward on what they learned and what they would do differently. And when you genuinely cannot make a coaching session, reschedule it rather than dropping it, because the reschedule says it mattered and the silent cancellation says it did not.

Where your judgment stays human

AI made Daniel faster at the structure, but every important call remained his. Is the person genuinely interested, or is this his goal projected onto them? Is the timeline realistic, or is he rushing? Can he actually show up consistently, or will he vanish when the queue spikes? Is this stretch developmental or just overwhelming? Is he offering enough support? And the quiet one underneath all of it: does he genuinely believe this person can do it? People can feel the answer to that question whether you say it or not. Fake belief does not work. Either you believe, or you are honest with yourself about why you do not.

Practice and reflection

Pick one person on your team and work through these. They are quick, and they tend to expose the gap between the plan you think you have and the one you actually have.

• Genuine interest check. For whoever you are currently developing, do they actually want this growth? Have you asked them directly, in those words? What is their real motivation, as opposed to the one you have assumed?
• Ambition reality check. Look at a development plan you have written. Is the timeline honest? What is genuinely achievable in one quarter, and what have you loaded in that belongs to next year?
• Stretch design. What would a good stretch assignment be for the person you are developing? Hard enough that it requires growth rather than just competent execution, and supported enough that they have a real chance of succeeding?
• Consistency audit. How reliably have you shown up for coaching conversations over the last three months? Count the ones you actually held against the ones you scheduled. If there is a gap, what keeps winning instead, and is it really more important?
• Belief check. For the person you are coaching, do you genuinely believe they can do this? Answer honestly and privately. If the answer is no, the coaching will not work, and you need to deal with that fact rather than perform around it.
• Real-world application. Choose one person to develop starting this week. Name the single most important growth area, set a realistic timeline, and identify the first stretch assignment. Then put the first coaching conversation on the calendar before you close the document.
• Reflect on the past week. Think of one conversation, decision, or missed opportunity from the last seven days where these ideas would have changed your approach. What would you have done differently, and what would the outcome have been? Writing it down is where the concept becomes practice.

Related lessons

Preparing Performance Conversations covers the conversation that usually sits just before development planning. Performance conversations tell someone honestly where they stand, which is the raw material a development plan is built from; without that clarity, the plan is guesswork.

Team Dynamics and Engagement widens the lens from one person to the group. Individual development plans work better when the team around them is healthy, and some growth blockers turn out to be dynamics rather than skills.

Feedback Crafting goes deep on the specific techniques for delivering the feedback that coaching depends on. Development stalls when feedback is vague or avoided, so the craft of saying the hard thing well is a direct multiplier on everything in this lesson.

Key Takeaways

• Development and performance are different things. Performance is the job someone does today; development is who they can become. Strong performers may have no growth appetite, and people with current gaps may have the most potential - look at trajectory, not just the current scoreboard.
• Five conditions drive growth. People develop when they see why it matters, get real stretch, receive honest feedback, have support, and keep agency. When development stalls, one of those five is usually missing.
• Development is a partnership. You owe vision, stretch, feedback, and support; they owe initiative, absorption of feedback, and applied learning. You cannot want someone's growth more than they do.
• Use AI to structure, never to coach. AI removes the blank-page problem and drafts a usable plan in minutes, but it cannot believe in your team member or build trust. The 20 percent only you know is the part that makes the plan real.
• Pace the plan deliberately. One major development focus per quarter, with an early achievable win for momentum. Going too fast crushes confidence; most managers err toward too fast.
• GROW plus 70-20-10 is a strong combination. GROW (Goal, Reality, Options, Will) structures the coaching conversation so the plan belongs to the person; 70-20-10 weights their learning time toward real challenging work, then learning from others, then formal training.
• Make milestones and support concrete and measured. Anchor progress to a baseline number, set dated checkpoints, and put recurring coaching sessions on the calendar as non-negotiable meetings.
• Permission to fail is what makes stretch possible. When someone hits a rough patch, normalize it and frame the mistake as information, not a verdict. Stretch without safety just produces fear.
• Name progress out loud. People cannot see their own growth from the inside, and confidence drains when improvement goes unmentioned. Coaching is a learnable skill, and so is noticing.
• Consistency beats brilliance. The plan works because you show up to every coaching conversation. Skipping them tells the person their growth does not matter, no matter what the plan says.