Let's pretend that I'm coming to you from the future. Your succession plan has names on the board, readiness ratings, green boxes even, development plans, and they're all on schedule. But here's the problem: the roles where these people would have learned to actually run something — they no longer exist, and they haven't for a while. You removed them three years ago in the name of AI efficiency, and your dashboard never noticed the lingering problem.
Hey there, senior leader, and welcome to the Talent Sherpa Podcast. This is where senior leaders come to rethink how human capital really works. I'm your host, Jackson Lynch, and today we're going to be talking about what flattening did to your leadership pipeline. And here's the really uncomfortable part: flattening, by all objective measures, worked. Decision speed went up, overhead came down, the business case was clean, and every consulting deck confirmed it. Nobody got fired at all for removing layers. But there was a second-order cost that never quite made it into the model. And it's sitting right now in your succession slate, disguised as readiness.
If you've ever sat in a talent review where the bench looked full on the slide but felt thin in your gut, this episode is going to name the mechanism behind that feeling. And with boards now planning workforce reductions of up to 20% based upon AI logic — which isn't very logical — this is about ready to happen again at a larger scale and faster. So by the end of this episode, you're going to have a way to rerun the succession math over a longer time period, and before the next wave of cuts locks in the lasting damage.
But before we get into it, I want to say a quick thank you. Shout out this week to Kay from North Carolina. Kay, I want to tell you how much I appreciate you being a part of this community. And to everyone that's tuning in, whether you're from Saratov, Russia or Altoona, Pennsylvania, I appreciate you being here. And to our friends in Russia — kak dela, ochen khorosho. All right, let's get into this.
I want you to picture a quarterly talent review. Your HR business partner is looking sharp. They're at the front of the room, they're walking through the succession slate for the top three roles at a very critical function. They have names and readiness ratings, development timelines, maybe a Post-it note and a couple colors. Everything is on track the way it always is. And then a board member says something that lands differently than it should. They ask the question: who on this list can actually run anything?
Now you look at the names again. Most were flagged as high potential two or three years ago. Several completed your very well recognized leadership development program. A few rotated through special projects — credentials, endorsements, completed curricula, and all of it is real. And the question, however, still hangs in the air, because the honest answer is that very few of them have led through genuine ambiguity with real consequences attached to the outcome. And that's what they need in the role we're talking about now.
The roles where that used to happen — that messy middle where the future VP learned to manage with incomplete authority, and a future general manager learned to balance competing priorities across functions — well, unfortunately, those roles got eliminated. At the time, we called it flattening, and we celebrated the savings. We talked about succession plans as board theater a few months back, and the episode was on why your succession plan is probably a lie. If you missed that one, it's worth going back. But today is the structural sequel to that. The plan is a lie, partially because the architecture that used to make it true is gone. So let me make this concrete.
I was working with a client recently who was being squeezed for additional cost-out. And spans and layers was the mechanism, as it has been since time eternal. Widen the spans, remove a layer, harvest the savings — standard playbook. To be honest, I've run that myself numerous times. And I told them there were two things to be concerned with, and neither one was the flattening itself.
First, the span math has genuinely changed. The more AI that you use today to do the work managers used to do, especially around cascading of clarity and evaluation of work product, the wider your spans can legitimately go. And no one really knows what those outer bounds look like. The old rules were built for a world where managers spent their days routing information and checking work. And that world is ending.
Second — and this is the fork-in-the-road one — the companies that thrive in an AI world will use that freed-up capacity to reinvest in the opportunity cost that they've been paying, in the work that they never had the bandwidth to get to. The growth work, the innovation work. And the companies that struggle will use AI as a machete against cost and then stop there. The machete companies are also, without even noticing, deleting every proving ground where their next generation of leaders is going to be built.
The savings show up this quarter, the hole shows up in three to five years. And in a talent review, when a board member asks the question: who on this list has actually run anything?
So the first trap that we're running into right now, I think, is measuring readiness as if the old architecture still exists. Your readiness ratings were calibrated in a world where people climbed through roles that tested them. Those ratings kept their labels long after the roles disappeared. So the slate says ready in two years, but the two years no longer contained the experiences that made the phrase actually mean something, if it ever really did. The metric survived the system it was measuring.
And here's where that usually breaks down for the CHRO or HR folks specifically. The org redesign arrives framed as a cost exercise, and the CHRO accepts that frame. So spans, structures, and savings get modeled to the decimal, maybe even a couple to the right of the decimal. The capability question — which one of these roles is producing our future operators? — well, unfortunately, that rarely gets forced into the room. Sound org design starts with strategic priorities and maps the critical decisions, assigns decision ownership, and only then does it draw the boxes. And one of the things that you would want to solve for as a strategic priority is building that next generation. Most flattening exercises, they go straight to the boxes. And when you remove a layer without mapping what capability it was producing, you'll find out what you deleted — and that is only going to show up when the absence of it arrives.
And the one thing that trips us up here, I think, as well, is the fuse is pretty long. Current senior leaders, they're still in place. The slates have names, the green boxes still render. The cost of a hollowed pipeline arrives down the road, diffuse and unattributed, while the savings arrived immediately and had an owner who got promoted for delivering just those dollars. Meanwhile, the pipeline is thinning from the other side too. Deloitte found that only 6% of Gen Z and millennial workers even name a leadership role as their primary career goal. Fewer proving grounds, and fewer volunteers for the ones that remain.
So here's the shift I'd like to offer you. Stop thinking about the middle layer as a supervision expense. That layer was doing two jobs at once. One was oversight — the routing and the checking, the clarifying that filled a manager's calendar. The other was development — the slow pressure-testing of judgment that happens when somebody owns an outcome that can actually fail. AI genuinely replaces much of that first function, and I'm all for it. That part of the flattening thesis is right, and it's getting more right every quarter. The old span-of-control orthodoxy, I think, probably even understates how wide spans can go now, because AI is absorbing exactly the work that used to cap them — the clarity work, the evaluation work, the status-chasing work. Maybe now we can actually give some employees feedback. That'd be awesome.
But when your cost-out or design model priced it in, that layer priced in the supervision function only. The development function rode along invisibly — unbudgeted, unmeasured — the way it always has. So when the layer disappears, the development went with it, for free, off the books. Nobody decided to stop producing leaders. Not intentionally, at least. It was a routing error in somebody else's business case.
That's the reframe. Spans and layers have been priced as a cost decision, and it is actually, at least partially, a capability production decision. Your org chart was also your leadership factory. When the layers come out, the factory shuts down. And the succession slate keeps reporting inventory as if the production line were still running. You're looking at an architecture problem that has been filed as a development problem.
Training was never what built those leaders in the first place. Consequence built them. Making mistakes built them. Having to stand up and explain your logic and your decisions made them. Standing up in front of your team — either smaller teams or at the all-hands — like, that's what built them. And once you see it that way, the question changes from "how do we develop our high potentials?" to "where exactly in this flatter organization does anyone really get tested?"
And that question has an answer, and it connects back to the fork I described earlier. The reinvestment companies, they have a structural advantage that no one is naming today. The new growth work, the innovation work that AI capacity makes possible, is the new proving ground. It's ambiguous, it's consequential, and it has no incumbent. And in fact, it can be staffed with the exact people that your succession plan claims are ready now.
So what's the playbook? Here's number one. Run the board member's question against your own slate. Take your top three succession candidates for your function and ask, for each name, what does the person really run? Run means that they owned a budget, owned a team, and had a way to visibly fail. Score every name on that single dimension and recount your bench. Bring the recount, if you can, to your CEO before a board member does the math for you. And the business value is — like, that's straight risk management. You're replacing a fictional readiness number with a real one, and every downstream succession decision is going to get way more honest as a result of it, the day you do it.
Play number two: force the talent math into the next org design conversation before the boxes get dropped. When the cost-out request comes — and let me be clear, it's coming — walk in with the sequence. Strategic priorities first, then the critical decisions, then the decision ownership, then the structure, and then the cost. For every layer on the cut list, name the capability it was producing and where that production now moves. Sometimes the answer is nowhere, and the business can accept that trade knowingly. That's fine. I'm not trying to tell you what the right answer is, only that you should make them with your eyes wide open. The point is it becomes a decision instead of an accident. And the CHRO becomes the person who made the real cost of these decisions visible.
Play three: convert AI capacity into proving grounds deliberately. Take the manager hours that AI is freeing up and tell them what to do with it. I would suggest point them at the growth and innovation work that your company never had the bandwidth for. Then staff those initiatives with succession candidates and give each one of them a real mandate — an outcome, a resource envelope, and consequences that show up somewhere visible. This is how you restart leadership production inside a flatter structure. And it pays twice, because the same move that rebuilds your bench is the move that also separates the reinvestment companies from the machete companies.
And play four: split your span model into two lines. For every manager role, set supervision capacity and development capacity separately. Supervision capacity expands as AI absorbs clarity and evaluation work — so it's okay, we should let those spans widen. Development capacity does not expand, because pressure-testing judgment still takes human attention in real stakes. For the handful of those roles that feed your pipeline, treat them separately. Set the span based on the development line. One sentence in your org design standards will fix this: spans are set by what the role must produce, and some specific roles must produce leaders.
So that's the slate audit, a redesign of how your company does org design, a leadership factory rebuild, and a new span model. It's just one episode, I know, and I'm aware of what I just did to the rest of your week. And I would lovingly do it again. And if you only take one thing from today, let it be this: your org chart was also your leadership factory, and you just flattened it. Rebuilding the proving grounds now is the succession plan.
So thanks for spending some time with me today. And I do appreciate you being a part of this community of senior leaders who want to rethink how human capital really works. Now, if you're thinking about how to apply this in your own situation, let me point you to a couple of resources. PropulsionAI is workforce intelligence for private equity. And that's it for today. Until next time, keep raising the bar — for real this time — and keep on playing.
