AI Workforce Reset: Reskilling, Redeployment and Leading Through Change

Introduction

Nolan: Hello everyone and welcome to the Talent Equation podcast. This episode is sponsored by Infopro Learning. As always, I’m your host, Nolan Hout. Today we’ve got Sean Bartman joining. If you’ve ever worked with somebody who asks why and why not and what if until your whiteboard runs out of room, you’re going to love this conversation. Sean has spent over two decades shaping talent inside some of the largest organizations in the world, including TD Bank, RBC, and KPMG, and he’s made a career out of poking at the status quo on behalf of the people inside of it.

Today, we’re going to talk about AI and the workforce reset, specifically how HR translates AI hype into redesigned work, reskilling, redeployment, and leaders who are ready for any of that along the way. With that, welcome to the podcast, Sean.

Sean: Hey, thanks very much, Nolan. So happy to be here.

Sean’s Origin Story in HR

Nolan: So how we start everything, Sean, is just understanding a little bit about who you are and where you are today. Obviously, you’re a senior leader at a large company, but you didn’t start there. What was your origin story into this field of working with people, developing people, and working within HR ecosystems?

Sean: Yeah, that’s a great question. I took a course at university. I was taking political science and philosophy, and at the end of my university run I took this one elective in human resources, and all a sudden the light bulb went on. I became that kid at the front of the class listening and answering all the questions. I realized I had just found an area of passion that I just wanted to chase. That started it off, and since then I’ve worked in tech, banking, and consulting.

I love every aspect of it. I’m hungry to work with great organizations and great people. I have a passion for working with particularly young and upcoming talents who just want to take on the world. I love to surround myself with that mindset. How do we get one step better every day? HR and this space, with AI lately, is just delivering that in an amazing way.

Nolan: And not to put you on the therapist chair for too long, but why do you think it was a passion for you? What made you latch on so quickly?

Sean: HR is a misunderstood word. When people say you work in HR, at a dinner table, you can see the reaction. But the more the conversation gets going, they start to realize that there’s coaching, career advice, learning, and they realize the HR space is massive. You can have ten, hundred careers within HR. And the ability to impact an organization through its talent is enormous. I think the mindset shift is that HR and talent should be viewed as an investment.

It’s the talent equation. It’s the talent side of running a business. When you start to shift that mindset from just doing work to how do we help run the talent side of the business, it just opens a world. I’ve been spending my time in as many of the domains as I can get into, and I’m loving all of it.

Forecasting AI’s Impact on the Workforce

Nolan: I think we forget that traditionally your spend on human resources outweighs your spend on almost anything else in your company. These are the people in charge of maximizing that return on investment. Right now, there are a lot of people drowning in AI, specifically as it relates to forecasting what AI is going to do to their organization and what the impact is going to be. The goalpost keeps moving. What’s your thought on that? How are you helping better forecast the work and the ROI using AI in HR?

Sean: That is a big question. First of all, it is really hard to predict the future of the labor market in general. AI is moving so fast we can’t keep up. Most organizations are still in the early majority, just getting their hands on it and putting it into practice. What my team is trying to do is predict where an organization would go in the future. What is the shape of the organization and what is the impact?

We do that through a lot of research, a lot of experience. In the banking environment, we have multiple pillars, revenue-generating businesses and supporting functions, that will be impacted differently. So, we step back and ask: what does good research say? What does good design theory say? Where might we predict the shape of the organization in the future?

The Pyramid to Diamond: How Organizations’ Structure Is Changing

Sean: Some functions are impacted differently. Knowledge workers are closer to what I’d call a diamond-shaped organization today. If you think of the organization as a traditional pyramid, we think that pyramid will slowly change over time into more of a diamond shape, as AI comes in and automates some of the work, accelerates it, enables it.

The knowledge worker functions operate closer to the diamond shape already. The impact AI will have there versus a revenue-generating business, which may be more of the traditional pyramid, will be different. This is what we’re trying to unlock.

Nolan: Can you double-click into the triangle? When you say most organizations are shaped like a triangle today, does that mean largest group of employees at the bottom, a few executives at the top?

Sean: Yes. The traditional pyramid has a CEO or leader at the top, a C-suite, and then more roles the further down you go. The concept of the diamond-shaped organization is that as you insert large language models, agentic AI, bespoke custom agents, you start to break down some of the work. The “got to do” work starts to be enabled, sometimes automated or accelerated by AI, and that cuts out some of what is traditionally the deep subject matter expertise and operational work at the base. That then starts to shift the organization from a pyramid toward more of a diamond shape. This is the really big rethink we need to tackle now.

Nolan: For me, it’s a really good visualization for a problem that’s hard to name. Mid-level management is one of the hardest things to define right now. Many of times those managers are the ones being made responsible for automating the bottom of the triangle. They’re becoming player-coaches. As we migrate toward the diamond, we’ll have fewer resources to manage. Are you seeing that mid-level manager becomes a more critical point in organizations moving through this shift?

Rethinking What It Means to Be a People Manager

Sean: Yes. We need to rethink what it means to be a people manager and what it means to be a leader as the organization evolves. This isn’t going to happen overnight. It’s going to be an evolution, as it is with any innovation and new technology. But if I break it down, I think we need to go from understanding roles to understanding tasks and subtasks. What is the actual work? From a bottom-up view, when we understand the work at the task and subtask level, we can re-bundle it. This is the type of work that can be done by AI. This is the type of work that can be accelerated by AI.

This is the type of work that is human. There may be some new work that comes out too. When you re-bundle that work back up, you get an evolved role for AI and an evolved role for people. Above that, the leader is now managing an integrated team of human and AI. The shift for the people manager is away from managing the task and toward understanding the value chain, the flow of process and work through AI and people, to deliver on an outcome. That’s the mindset shift of the new manager.

Nolan: So, what you’re really chasing is context. These managers need to have business context for what they’re doing. They need to understand where the ship is headed, how they’re being measured, what their role has to do with the performance of the overall organization. The skills and tasks will change over time, but if they can always anchor to how they’re being measured for performance and what their impact is, that’s something more durable. Is that what I’m hearing?

Sean: Yeah, I think so. You still must manage, coach, and develop people, but what you’re managing and coaching them for is different. It’s less about the task and more about the decision-making process, values, culture, risk management, creative problem solving. Those are where we need to invest in how we develop our leaders and people. They still have business outcomes they’ll be measured against. It’s just how they go about creating values that will be different.

How TD Is Approaching the Transition

Nolan: So how do you, within HR and talent development, shape this transition? It’s a phenomenal opportunity but also incredibly difficult. How are you starting?

Sean: We’re starting by elevating the talent within our own team and embracing AI and technology for ourselves. Put on the oxygen mask first, make sure we understand it and are integrating it. But I think the key thing right now is that a lot of people across the organization are running and doing things with AI in their own verticals, and we need to now move broadly. Most organizations got excited, and now we need to become more integrated. Thinking only about the business outcome is insufficient. Thinking only about AI or technology is insufficient. Thinking only about the process is insufficient. Thinking only about the people is insufficient. These groups need to come together to ask: what is the business outcome or client experience we’re looking to drive? What is the flow of change? And what is each group’s element in that change? Then, from a people perspective, how do we break down or rebuild work, workflows, roles, teams, and structures to enable that?

There is also a broader enterprise investment case: raising the tide of AI capability across the organization, investing in leadership savviness around AI. But as you execute, the importance of coming together will show success. If people continue to work in silos because they’re excited about something, that will work for the short term. But I think organizations will trip.

Breaking Down Silos: Why Integration Is the Unlock

Nolan: We’re seeing this too. We put out research a couple of months ago showing that AI is having significant value in everything we do, but there’s almost a multiplier that happens when you integrate AI-driven solutions together. If you’re just leveraging AI sitting on top of your learning management system, you only have the context of that person in the LMS.

You have no idea what they look like from a performance standpoint, what their 360 results show, how long they’ve been in their role, what their career path looks like. The better you can gather more context, not only does it make you better, but from a structural standpoint you’re giving AI more context. Those of us who have been working in it know how important context is. The more context you can give it, the better it performs.

Skills-Based Organizations and Real-Time Skill Inference

Nolan: You mentioned reskilling. How important are you seeing the specific skills of individuals in this whole future of work picture?

Sean: Incredibly important. We are on that journey as well as a skills-based organization. We have capability that does skills inference roles and colleagues, to say here are the skills that exist and here is the demand for skills across the organization today. As roles change and evolve, it automatically identifies how the skills needed are shifting. From a colleague perspective, we can also understand the skill supply we have.

We can do targeted assessments to say, here is a specific job population, here is where we are strong and here is where we are light, and that informs talent development. We can cut the data many ways. We have our finger on the pulse of supply and demand in a dynamic, up-to-date way.

We can take targeted action and macro action and arguably predict where we’re going to go. This is still early for us, but it feeds into the talent marketplace, talent management, talent acquisition, learning and development. And it’s beneficial to the colleagues themselves to plan their own career and skill development.

Nolan: The progress in skills over the past two years has been significant. If you’re not in the field, it’s hard to describe. Think about how you used to try to measure the skills of an organization. First, you had to define what skills to measure. Then HR or L&D would go to a business head and ask, what skills does your marketer need? And there was no common language to translate between the HR perspective and the business perspective. You’d finally agree on twenty skills, and then by the time you’d done the gap analysis and figured out how to close it, for organizations of the scale you work with, it was years. By the time you finished, it was out of date.

With AI, you can now do a pretty good job of helping define what those skills should be for an organization and do a solid initial job of validating them. Where a lot of companies are focusing now is on closing the gap. But I always tell people, just because you don’t have a way to close the problem yet doesn’t mean you shouldn’t invest in identifying it first. You have to do that before anything else. Don’t let the absence of a closure strategy slow you down. Everyone is at a different stage of this journey. Getting there is quicker than you think.

Predicting Future Skills in a Fast-Moving Market

Nolan: One of the things that comes to mind when we talk about reskilling in the age of AI is that we’re trying to predict the future, which nobody can. We’re trying to predict the future skills someone will need, while a lot of the skills people are using today may be the ones that get carved out.

How do you define what those future skills are, and how do you get people to invest in developing skills that, in their own minds, might feel like they’re accelerating their own replacement?

Sean: The right question is, what are the skills? And the answer is that skills are continually evolving. So, the goal is to move from guessing the skills to greater certainty about the evolving skills. Using AI to understand the moving market helps with that. But practically, you then need to do something with that. It can’t be changing constantly. You have to act. Where the move with AI takes us is that tactical, task-level work will be handled increasingly by AI. So, are we going to train people on data entry? Probably not. Where we need to shift is toward relationship management, risk understanding, AI auditing, knowing where to look and how to evaluate outputs.

Creative problem solving especially. The human side is: we now have this capability and this data. What is the art of the possible? And we have better insight than ever. Through employee listening, through skills data, we can query the workforce and get real answers. We had guessing in the past. Now we have greater certainty about what’s important to the colleague base and where to invest. Everyone needs to shift away from the “got to do” work and toward ownership of outcomes. That’s the mindset shift that takes time, but it’s where we need to invest.

Shifting from Task Execution to Creative Problem Solving

Nolan: I really like what you said about creative problem solving, because I think a lot of people assume AI is for tech people, and the more technical you are the better you can use it. The problem is that AI gives you the ability to do almost anything. So, the question is no longer what can be done. It’s what should be done. Understanding what should be done is genuinely hard.

In the L&D world, I can technically now create unlimited amounts of eLearning because what took a day takes an hour. But does that mean I should create the content eight times? No, because the content that is already out there isn’t being consumed. That curiosity, creativity, critical thinking, whatever you want to call it, paired with business context, is going to be a critical skill set.

Sean: Exactly. And that’s triggering something I want to name. We’re at the point where, as people are getting excited, the question becomes: are we creating value with AI, or are we creating burden, at pace? You can churn more output, but are you just filling up someone else’s inbox that they’re then going to use AI to process? Are we doing work for each other, or are we doing something someone is going to say thank you for?

Are We Creating Value or Creating Burden with AI?

Nolan: Yes. A very clear example: it used to be that if I wanted to craft an email, it took me longer to write it than it did to read it. Now it takes longer to read it than to write it. If you are putting a one-sentence prompt into Claude and then sending a seven-page memo to your colleagues saying, here, read this and tell me what you think, you haven’t earned that.

You haven’t invested enough in it. You’re shifting the burden onto other people to disprove whatever you’ve said. I tell people you should be investing as much time creating the solution as you would expect someone to spend reviewing it.

Redeployment: Where Do Reskilled People Go?

Nolan: Let’s assume we’ve reskilled people. We’ve identified the skills gaps, we know the future skills, we’ve given people the development. But how do we know where to put them? A lot of organizations are restructuring right now, asking whether they even need the same headcount. How do you know where to redeploy people once they’re reskilled?

Sean: Back in the day, there were artisans making quilts, and then they introduced the loom. People still make quilts. They learned how to use the loom. Then we had people using pen and paper to do calculations, and we introduced the computer. People had to learn to use the computer, and then we had new roles to build computers and service them. In banking, we got the ATM and then online banking.

The frontline teller shifted more into relationships, sales, and service. The type of work changes when the “got to do” work goes away. It moves to what I call the “want to do” work, or the higher-impact work. And in many cases new work comes out entirely. Ultimately organizations have a choice: what do you do with this newfound capacity and capability? Do you redeploy it for growth?

For better client experience? For innovation? Do you focus on cost reduction and pocket some efficiency? That will happen. But the way organizations outpace, and win is not to focus only on the takeaways. Think about what’s the investment. How do I invest to win differently? That is the key right now.

Nolan: And a good example of that is financial institutions. When ATMs came in, you could have just cut everybody and moved fully online. That saves money, a lot of it, right away. But the financial institution I use offers notary services now. They realize that if someone has their money there, when they must make big decisions about their will or sign a lease, they need a notary.

By offering that service, they keep me as a client. If you race to the bottom, unless you are the absolute best at being the cheapest, like Walmart, you’re not going to win. Second cheapest doesn’t cut it.

Sean: Yes. Where the gain is, is taking complexity out of the organization. Increasing speed. Executing with excellence and discipline. Creating value, not burden. Show me any organization not looking to grow and have an impact. The one next to them is going to pass them pretty quickly if they take advantage of this opportunity. As individuals and as organizations, this is a fantastic opportunity in front of us right now.

Educating Leaders on How to Use AI-Freed Capacity

Nolan: This is specifically where educating our leaders matters most. What is the mindset, what is the culture do you want these employees to carry through their ranks? You can have one manager who says, you can now do your job ten times faster, so I want ten times the work. Or you can have a different manager who says, I know it doesn’t take you the same amount of time anymore, and that’s okay. What I want you to do is find ways to add more value somewhere else.

I remember having this conversation with one of my content developers. Content development is an area where AI has really helped us accelerate. What I said to them was, I need you to focus less on what you’re developing and more on the impact. Help me understand what content has the largest effect on learning. What are the shifts and trends?

What type of content does people want to see? What are new ways we could service content? Because AI is just going to keep making production easier. What AI can’t do is help us make better decisions about what to produce. And if you’re not an organization focused on cutting costs, you should be having those conversations with your leaders about how to handle this excess capacity. Finally give your people time to be strategic. We’ve been stacking increasingly on them for years. Let them breathe a little.

Sean: Where people and organizations trip is this: as you progress in an organization, you spend less time doing and more time leading and being strategic. But as people get senior, they don’t always let go of the doing. They get stuck going down and into the details. What AI is doing is freeing leaders to rise up. Instead of spending time looking down and in at task management, they should be looking up and out and forward.

With the freedom of not having to manage every task, trust your people, trust the integrated workforce, and start lifting your gaze. Look down the road, maybe even to the horizon. What is the art of the possible with this team, with these individuals? Plan forward so that the next level can execute and deliver. We get caught looking at our toes when we should be looking up.

Build Meta Skills Before AI Tool Skills

Nolan: And we even did that as an industry, right? The second AI came out, everybody said, let’s just go teach everybody how to use AI. But we didn’t stop thinking. The companies that saw the bigger payoff were not the ones that said, let’s implement Claude training. They were the ones that said, let’s go really build critical thinking. What does curiosity mean? What does critical thinking mean?

And how do you apply that in an age of AI? There was a short-term loss because maybe their people didn’t know how to write a better prompt. But those gaps were quickly closed when they had these meta skills developed underneath. As we close, I want to ask you one last question. Five years from now, we’re looking back. What do you think we’ve overreacted to? And what do you think we’ve underreacted to?

Five Years from Now: What Did We Get Wrong?

Sean: I think things will be completely different. We’ll see if the prediction of the diamond truly comes or if it’s just a shorter pyramid. Maybe some people disagree that we’ll ever get to a diamond shape. But I think what will happen is we talked a lot today about managing risk, auditing AI, being the human in the loop. After a while, when we’ve checked it and it’s right again and again, we’re not going to need to keep checking it at the same rate.

That will release more capacity and shift the mindset further. I think we’re going to see the nature of work and how people think about work fundamentally shift in five years. People will start thinking not just about what can I do, but what is the art of the possible. Where we have it wrong right now, I’m not sure. If I knew, we wouldn’t be doing it. We’ll see.

Nolan: Do you feel AI is overhyped or underhyped?

Sean: I’m the guy who is passionate about it. It’s integrated into my life. I think it’s underhyped right now. There are a few people at the forefront, the early adopters and fast followers, who really understand it. But a lot of people are still just starting to realize. I had a friend who said they didn’t want to engage with the robot. I said, okay, let’s just try something simple. Let’s create a change management plan for this topic with these constraints. It spits out this really detailed, genuinely good change management plan. And all of a sudden, they went from being a naysayer to being in it and using it every day. As people have these aha moments, we’re going to see adoption accelerate.

We’re still at the surface of basic prompting. As people start to experience advanced prompting, see their processes accelerated or automated, and as agentic AI starts to shape more of what’s possible, when the average person starts to realize what tools like Claude Code can do, there’s so much untapped that is going to blow people’s minds in a good way as they integrate it. People will do that at home first and it’ll come to work a little slower. But we are just at the tip of this.

The Ethical Dimension We Are Underweighting

Nolan: I’m all in on underhyped. But not just from excitement. I also feel there is some impending risk. If I’m being honest, I think we’re under hyping how powerful it is, and therefore also under hyping what it could do in the wrong hands.

In history, sometimes we have to have the bad to recognize and strengthen the good. I hope AI somehow helps us prevent the bad. I hope the good guys stay one step ahead.

Sean: I hope that too. But you’re right to raise it. We talked about the other skills and capabilities needed, and the ethical side is one of them. It’s not just about checking AI to go faster. It’s also about asking: is this the right thing? Is this doing good, or is this to the detriment of whatever outcome we’re trying to reach?

That is important to weave into organizations. What are our values? What is our culture? What is our ethical line on this? And how do we protect against those who don’t share those values? The defenders of that will be a critical new role. As much as we both love to embrace AI, I have the occasional pause too. But I’m still on the side of moving forward.

Nolan: Well Sean, thank you so much for giving us this time today. We really appreciate it and we hope maybe we’ll have you on again soon.

Sean: Thank you, Nolan. I appreciate it.