Know what you have
TL;DR. Most organizations run workforce planning reactively or on a calendar. Neither approach sees what's already inside. Capability-led planning flips that — you understand what your people can actually do before you go hire for what you don't have. The hard part isn't the idea. It's scale, trust, and managers willing to let go.
Most workforce plans start the same way. Leadership sets a strategy. Finance builds a model. HR gets headcount targets. Recruiting starts filling seats.
It works until it doesn't.
The strategy assumes the talent exists, that you can hire fast enough, that the skills you need are available in the market. Sometimes all three are true. A lot of times, they're not. And by the time you find out, you're already behind.
There's a different way to think about this. It starts by looking inward before you look outward.
From reactive to predictive: a quick detour into heavy equipment
Bear with me for a second.
The way organizations manage large equipment has gone through three distinct phases.
The first was reactive. Something breaks, you fix it. Simple. Also expensive, unpredictable, and disruptive, especially when the failure happens mid-job on a multi-million dollar project.
The second was scheduled maintenance. You don't wait for the breakdown. You pull equipment out of rotation at regular intervals, whether it needs it or not. More predictable. Fewer emergencies. But inherently conservative. You're over-maintaining some equipment and under-maintaining others. You're using a calendar when you should be using data.
The third is where things get interesting: predictive maintenance. Sensors on the equipment. Real-time diagnostics. Pattern recognition across thousands of machines over years of operation. The system tells you, with surprising accuracy, when something is likely to fail before it does. You intervene at the right time, not too early and not too late.
But here's what makes modern predictive maintenance actually work: it's not just the data. It's the people who run the equipment every day. The operator who knows that particular machine vibrates differently when it's hot. The mechanic who recognizes a sound before the sensor does. The data and the human knowledge work together. Neither one is enough on its own.
That last part matters a lot for where this is going.
Where most workforce planning still lives
Most organizations are somewhere between reactive and scheduled.
Reactive looks like: a key person leaves, a project stalls, a hiring surge kicks off. The plan was fine until it wasn't.
Scheduled looks like: annual headcount cycles, tied to budget seasons, built around org charts and job families. Better than reactive. Still slow. Still built around roles and headcount, not around what the organization can actually do.
Both approaches treat talent as something you go get from outside, not something you understand and activate from within.
The whiteboard version of this
When I took over a new team, one of the first things I did was get everyone in a room and ask them to be honest about themselves.
What are you genuinely good at? What are you still working on? What do you love doing? What drains you?
We didn't call it a skills inventory. That would have killed the mood. We called it getting to know each other. But that's exactly what it was.
What came out of it was a baseline. A shared map of where we had strength, where we had overlap, where we had gaps. It helped us divide work more intentionally. It helped people lean into what they were good at instead of defaulting to whoever was least busy. Over time, as I got to know people better, I could see where the initial self-assessment was off. Usually because people had undersold themselves.
It was manual. It was imperfect. It worked because there were five of us. I was not a VP with 250 people reporting to me. There was no system, no ongoing feed of information, no way to keep it current without constant one-on-ones and gut instinct.
That's the problem. The idea was right. The execution didn't scale.
What capability-led planning actually means
The shift I'm interested in is this: before you look at who you need to hire, understand what you already have.
Not job titles. Not tenure. Actual capabilities — the skills, patterns, and approaches that show up in how people work, not just what their résumé says.
The gap between "job title" and "what this person can actually do" is enormous in most organizations. And it creates real, concrete missed opportunities.
Consider this. Right now, somewhere in a large organization, there's a retail employee on the sales floor who is finishing a master's degree in data science, or organizational psychology, or supply chain. That person is in the system as a sales associate. The org doesn't know what they're building. They might not think to mention it. When a relevant role opens up, the organization goes to market, pays a recruiter, waits three months, and brings someone in from outside.
Hiring didn't fail. The org just couldn't see what it already had.
When you close that gap, a few things change. Strategy gets more honest, built on execution capacity that actually exists, not optimistic hiring assumptions. Hidden options emerge — the PM who knows how to build a data pipeline, the analyst who came up through operations, the designer with a background in behavioral economics. Talent investment gets more targeted. You're not sending people to training because it checks a box. You're building a specific capability you've identified as critical to something real.
People need to be at the center of this. Full stop.
I want to say something directly, because I keep running into this in strategic planning conversations.
There's a tendency to focus almost entirely on what this means for the business. The efficiency gains. The planning accuracy. The reduced hiring costs. Those outcomes are real and they matter. But if the conversation never gets to what this means for the people inside the organization, you've already made a mistake. And you'll pay for it in trust.
I've heard it directly: "HR is only asking us to update our skills so they know who they can lay off."
That's the narrative that takes hold when employees don't understand what's happening or why. Once it takes hold, it's very hard to walk back. People stop engaging. The data gets worse. The whole system becomes less useful.
The only way around this is to make the value real for employees, not just leadership.
Think about what it would mean to have a clear, evidence-based picture of your own capabilities. Not just what you think you're good at, but what your work actually shows. Most people underestimate themselves. Impostor syndrome is real. Self-assessment skews low. If a system could show you, with evidence, that you've been underselling yourself, that's not surveillance. That's a tool for your own career.
The goal isn't a company database built for workforce strategy decisions. It's a living picture of what people can do. One that's useful to them first, and useful to the organization as a result.
Managers are the bridge — and they have to let go a little
This is where the equipment analogy comes back around.
In predictive maintenance, the data and the operator work together. The sensor catches what the human misses. The human catches what the sensor can't measure. Neither one is running the show alone.
The same model applies here. At scale, you need systems and data to see across a large, complex workforce. But data alone won't tell you everything. Managers have context that no system can fully capture. That context needs to flow into the picture, not sit locked in someone's head.
Here's the problem: managers often hold on too tight.
They know who their strong performers are. They've invested in them. They want to keep them. So they don't raise their hand when a great opportunity comes up elsewhere in the org. They protect what they've built, even when the person on their team might thrive somewhere else, or when someone else in the org desperately needs exactly that skill set.
It cuts the other way too. A manager might highly value someone who is strategic, outgoing, and always thinking about the big picture, but undervalue the person who goes deep on one thing, executes with precision, and gets the hardest deliverables across the finish line. Both profiles matter. Just not always to the same team at the same time.
When capability data flows more openly, with managers actively contributing to it and using it, a few things get better. People get matched to work that fits them, not just the team they happened to land on. Managers who've been hoarding talent lose the ability to do so, because the visibility is there. Employees who are ready for something different, or whose skills are genuinely more valuable somewhere else, have a path to make that case. Not because HR decided it, but because the data supports it.
The manager isn't replaced in this model. They're essential to it. But they have to be willing to use the information in service of the person, not just in service of their own team's short-term needs.
The honest challenge: scale breaks this fast
Getting accurate skills data at scale is genuinely hard. I want to be clear about that, not sell past it.
Start with who you're even counting. Most large organizations aren't working with a clean, unified employee population. They have full-time employees in multiple countries, contractors who've been embedded for years, outsourced functions, augmented labor that sits somewhere between vendor and team member. Each category comes with different data access, different privacy rules, and different levels of organizational visibility. In some regions, the legal constraints on what you can track, store, or act on are significant, and they're changing. Building a skills picture across that landscape isn't a data problem. It's a data, legal, and organizational design problem all at once.
Layer on top of that the reality most large organizations operate in. Ingrained processes that took years to build. Org structures that have been through multiple transformations. Teams that have learned to work around systems rather than with them. You're not dropping capability-led planning into a clean environment. You're introducing it into one that's already moving.
Self-reported skills have obvious limits on their own. People overstate some things, understate others, and often don't know how to describe what they actually do in terms that match what the organization is looking for. Inferred skills are better in theory, but the models aren't perfect. They reflect what's visible digitally, which disadvantages people who do their best work in conversations, in rooms, and in relationships. Even when the data exists, job titles don't map cleanly to work, tenure doesn't predict capability, and the person with the most relevant experience for a new initiative might be buried three levels deep somewhere nobody is looking.
None of this makes this approach impossible. But accuracy is a journey, not a launch feature. You build it carefully, start with the right use cases, be honest with people about what the data can and can't tell you, and keep investing in the feedback loops that make it better over time.
What this looks like in practice
I've been working through versions of this at a large, global organization. Connecting workforce decisions to skills, business needs, and future demand across a complex, multi-segment employee base that spans employment types, geographies, and functions. Most large companies are still in early innings. The tooling is improving. The data models are getting better. The organizational muscle for this kind of planning is still being built.
But the direction is clear. The orgs that learn to see what their people can actually do will move faster and make better decisions with less guessing. And the people inside those orgs will have something most of them don't have now: a real, evidence-based picture of their own capability, and a path that reflects it.
That's the version of this worth building toward.
It's not complicated to describe. It is hard to do. But the starting point is the same whether you're working with a whiteboard and five people or a platform and fifty thousand.
Know what you have before you go looking for what you don't.
A note on how this was made: the ideas and experiences here are mine, built up over years of managing teams, building systems, and sitting in the planning sessions where these tensions actually play out. I used AI to help me structure and edit the writing. The thinking is mine. The polish had help.