Your AI Wrote Your Development Plan. So Whose Growth Is It? - Forbes
High confidence: full text extraction produced 6516 characters.
Why the struggle to name your own growth areas is the growth
Nikhil couldn't explain his own development areas.
We were on a video call, reviewing the homework that anchors every coaching engagement I run. After I interview a dozen or so of a client's stakeholders and synthesize what I hear into a feedback report, the client writes three leadership development areas: each one a headline problem statement in the first person, followed by a brief elaboration. The document Nikhil sent me was immaculate. Clean structure, confident prose, not a comma out of place.
Then I asked him to walk me through the first one. He stumbled. He circled. He glanced down, plainly reading his own document for orientation. Finally he confessed: AI had written it. All of it. He'd fed in his feedback report, accepted what came back, and moved on with his week.
The part that troubled me most was what came next on his calendar. In my process, those development areas go in front of the executive's manager at a checkpoint meeting, where the goal is to align on the growth agenda and get the manager emotionally invested in it. Nikhil was days away from presenting a document he barely recognized as his own.
The Most Ironic Delegation in Business
To be clear about what went wrong here: the problem wasn't the technology.
I'm a heavy AI user myself. It helps me summarize stakeholder interviews, draft session notes, make sense of contracts, and serve as a thought partner on hard problems. I even have AI generate a draft of a client's development areas from the feedback report I write, and it usually does a good job. But I position that draft as food for thought, a comparison point, never an answer key. Nikhil and I used the same tool on the same task. The difference is that mine was a conversation starter, and his was a substitute.
The purpose of writing your own development areas isn't to produce a document. It's to force a reckoning: taking a dozen stakeholders' worth of uncomfortable feedback and wrestling it into first-person sentences you actually own. “I tend to.” “My pattern is.” The document is just the receipt.
Outsource the writing and you skip the reckoning. The homework exists to build self-awareness. Handing it off defeats the entire purpose of the exercise.
The Struggle Was the Product
I know when a client has done the work themselves, because the first draft is usually kind of crappy. And I don't mean the grammar. The ownership language is missing, and in its place I find deflection, justification, and a little quiet blaming of circumstances. It reads like the early stages of grief, because that's roughly what it is.
That's healthy. A defensive first draft is diagnostic gold. It shows me exactly where the client is explaining away their feedback, and working through it together, iteration by iteration, is the growth. Most clients need two to four drafts before we land on something they'd be proud to put in front of their manager.
An AI-polished first draft skips that struggle and hides the evidence. The deflection that would have shown up in the client's own clumsy sentences never makes it onto the page, which means we can't work on it.
I see the same dynamic from another angle. When I share my AI-generated draft with a client who's struggling, most take it as intended. A few, deep in denial about their feedback, declare the AI flat out wrong. It usually isn't; it's read the same feedback report I wrote, and it's summarized the patterns faithfully. When an executive disputes a machine's neutral summary of a dozen consistent stakeholders, that tells me a lot about how far we still have to go.
Research is starting to catch up with what I see in coaching sessions. Preliminary MIT Media Lab research tracking brain activity found that people writing with AI assistance showed the weakest cognitive engagement, and most couldn't recall what they had just “written.” The researchers called the accumulating effect cognitive debt. A 2025 study of 319 knowledge workers by researchers at Microsoft and Carnegie Mellon found that the more confidence people reported in generative AI, the less critical thinking they applied to its output. These are early studies with small samples, so I hold them loosely. But point their findings at some of the most consequential writing an executive does, writing about what behaviors they need to improve, and the risk is super clear.
One nuance is worth pointing out. Some of my clients are brilliant technical leaders working in their second or third language, and for them AI is access, not laziness. I understand the appeal. I'd just note that the first draft is the worst place to take the help, because the first draft is where the growth lives.
A Helper, Not a Ghostwriter
I don't think the answer is an AI ban. I think it's knowing which part of the work you can hand off. With my clients, I recommend a sequence that looks like this:
- Draft ugly first. Your words, deflections and all, before AI touches anything.
- Then invite AI in as a sharp colleague. Let it tighten a near-final version, play devil’s advocate, and ask what you're not seeing.
- Finish with the test that caught Nikhil: put the document away and explain each development area from memory, out loud, as if your manager had just asked about it in the hallway.
That last step is the whole game, because polish was never the sin. The version Nikhil eventually brought me was still polished, and he may well have used AI along the way. The difference was that he knew exactly what each line actually meant. At the checkpoint, Nikhil walked his manager and a senior colleague he’d chosen to support his development through each area, and took their candid feedback without flinching too hard. I facilitated, but he carried the room. Before the meeting ended, that colleague told him he was impressed by how little ego Nikhil brought to a conversation like that. The work isn't finished, and really never is, but now it's truly his.
The executives who grow fastest over the next decade won't be the ones with the best-written development plans. They'll be the ones who keep doing the work that can't be delegated: wrestling hard with valuable feedback until it becomes the truth in their own words. The homework was never the document.
Your go-do this week: take one piece of hard feedback you've received this year and write three sentences about it. First person, no AI, no polish. Then read them out loud. That wobble you feel is the homework. Don't outsource it.