AI is everywhere. The agentic organization isn’t—yet

Just about everybody in the workforce is going to need a new job description in the next two three years. So most roles won't go away, most roles will actually be reshaped. That's McKinsey Senior Partner, Alexis Crv. COVID. She's talking about how Agenda AI will transform our jobs and she joins us to get into the ways leaders can reorganize around this consequential change. This is the McKinsey Podcast where we help you make sense out of the world's toughest business challenges. I'm your host for today, Lucia Reheli. Alexis Crv. COVID. Welcome to the podcast. Thanks for having me. Alexis, we hear much in the media about massive investment in Agenda AI without corresponding returns yet. And in fact, our own research shows the complexity of delivering on the promise of Agenda AI at any kind of scale. So let's start with some context on where we are now to get a grounding in the facts of the research. So the great paradox is just what you opened with. 80% of companies see massive transformation coming from AI and have invested with that mindset. And yet more than 80% of companies would say they're not yet seen the impact to the bottom line from those investments. So the real question is how does the organization meet this moment? How do we create the agentic organization of the future? And in every conversation I'm in companies feel like they're on the precipice of that. And what they're really trying to think through is a set of questions of how our roles going to change. What are the skills we will need in the future? How do I bring a workforce along with excitement, not fear? And really, how do I drive that change into an organization when it hits every corner of the company? In the research here in your co-authors outline five pillars of the new agentic organization. I wonder if you could just briefly set the stage for us with some examples of agentic for an organization's business model to sort of help illustrate the value that stake for leaders. One aspect of business model is obviously how in a world that could move towards near zero marginal cost of delivery. How does that change what you're able to bring to customers? How you're able to tailor and hypersegment to the unit of one and how you think about that in the context of mobilizing growth? What does it look like if you're a tech player that does content distribution and you can actually create a tailoring experience down to the individual consumer across millions and millions of people? What is that open up in terms of how you can kind of create and enable commerce? But then you take that a step further and the question becomes what if on the receiving end, what if the small business, the individual has agents of their own that are interacting on their behalf? What if I have someone that manages my financial accounts and now can move my money frictionlessly across bank account to bank account to seek best rate? That fundamentally changes the mode that has existed in financial services since the beginning of time that you don't do these things because deposits are sticky because the process of making these changes is so incredibly hard. But if that suddenly goes away, the entirety of how you think about your business model both in terms of opportunity, but also risk associated with competitive threat changes. At the broadest level, this is about how organizations respond when the assumptions underlying work begin to shift. The question then becomes what that looks like inside the organization itself. So Alexis, you are one of the global leaders of our people in organizational performance practice. Talk to us about how you see team structures functioning in this agentech era. The real promise in this moment with agentech relative to generative AI or previous evolutions of AI is the idea that you can have the equivalent of superhuman capabilities augmented and added to your teams. What the day-to-day workflows look like and the rituals around ways of working are going to need to change fundamentally. And that's I think what we mean when we say the operating model needs to shift. You need to think about how the hours of the day happen differently. The process of overseeing an agent population, how you engage as a team when you problem-solve around that and put the right governance and risk controls on top of that. And for most companies, this is all new ground. And we now have the opportunity and the challenge to grapple with it nearly everywhere simultaneously because the use cases are so broad. As you're noting, it's super early and everyone's model is being upended. But in your work with pioneers in this area, any sort of examples where are folks starting in terms of rethinking their org charts, for example? Well, org charts is a tricky one in particular because I think we're still too early to say I think the answer is going to look quite different domain by domain. Take pharma, a lot of life sciences and pharma companies are envisioning huge squads of agents in the R&D space. That doesn't mean they're not going to need the researchers and scientists they have today that they will be replaced in the org chart. It means they're going to be able to supercharge the speed at which they can do innovation. But in some other areas, companies really envision in corporate functions that are cost centers that deliver critical, important execution capabilities into companies, but they do not drive the front end of generation of value, HR, finance, legal. There's a real hope that what we can do with an agentic capability is actually transform how that work gets done, who does that work, and in some places to enable actually more capacity to go to other aspects of the business. Many companies have added at least one layer to their structure from CEO all the way to front line over the past decade. In some organizations, that's more like two or three. Not only is that expensive, but that slows companies down from a decision making standpoint because it just means you have more people, more layers at which somebody has to weigh in before any decision can get made. And there's a real hope in this moment that one of the things that AI can enable is actually speeding up those decisions and those connection points, allowing leaders to have even more of a superhuman capacity to manage across bigger scopes, which would allow companies to flatten their structure and get faster in the process. What should leaders be doing to develop the skills they need in order to manage both these broader scopes of talent and to have the skills they need to oversee agentic tech? Just about everybody in the workforce is going to need a new job description in the next two, three years. So most roles won't go away. Most roles will actually be reshaped. If you are sitting in the C-suite or the executive levels today, 70% of roles need fundamental reshaping already. Right now in this moment, and that includes the very people leading the teams who need to rethink their own job description as they rethink it for those that report up and into them. So when we ask about capability building needs, what we hear, no surprise, is nearly half of leaders say they think they see skill gaps in their organization and even in themselves. And most would say they really benefit from more training, more capability building, more support, and that goes all the way up into senior leadership. Does that also have implications for the organization of the tech function? How will the tech function change? I think there are a lot of different models for how that's playing out in companies. Many are deciding to create these sort of SWAT teams that go out and support that type of transformation, sort of draw on your tech teams and enable that. Others are embedding expertise inside each area of the business. But I think it is a classic example where you're both trying to change myself and transform the organization at the same time. It's more than just a change management journey. It's like a change management Odyssey, right? So how should leaders be thinking about leading that change? One big question for leaders in this moment is how do I need to show up differently in order to lead the organization forward? So I'll give you two examples. I know in one company where they said, if this is truly going to radically transform everything about our business, we as a leadership team need to start by radically transforming ourselves. So is 50% of my time spent differently because I now have access to AI to do my job or am I dabbling with it in the corners and in the pockets of tightening an email here asking a queer there, but not really rethinking the hours of my day. And then if I am radically changing how I spend my time, how do I talk about that in the organization? And then there's a huge trust gap that exists today, both because some of the early experimentation with AI has proven to have real faults, the hallucination stories that we read in the headlines. They're real, the slop that moves around a company as poor use or uncontrolled use of AI gets flung back and forth, can create more work, not less. So how do you as a leader break through that and get people optimistically excited about what's possible, but still with the overlay of good judgment and risk management to know that we're in a learning mode, we're not in already a mature state of deployment. Well, there's a element of fear also, right? Folks think they may lose their jobs and maybe replaceable by a genetic AI. There must also be a certain amount of resistance that's fear driven as well. For many companies, leaders need to think through what is the narrative that is genuine, but also creates positive energy around this change because the technology is not going to stop. And professor who talks about you won't be replaced by AI, but you may be replaced by someone who embraces AI before you do. And I think that's a real question out to the employee base. How do I get you excited that there's actually something here about new skills, new tooling that could make you more successful? And if it's not in this role, because this role will transform so dramatically, perhaps it's in another role adjacent. And on that topic, do you see any new talent profiles that you think will be essential? And how my employees begin to develop those skills? So some of the things that we're seeing go down, you'd expect maybe because AI is really good at this stuff. Research, skills, mathematical skills in some places, data science skills, but other aspects of those same skill profiles are becoming even more prominent. So the strategic thinking, the systems orientation, the soft skills of the people management, all of those are on the rise. And it's not to say that the first set of skills aren't needed anymore, but in an AI augmented world, you'll be able to do so much more of that with the assistance of agents and what we'll need even more of is the oversight, the judgment, the problem solving capabilities, the operational capabilities that human leaders can bring as the overlay to all of that capability. We have previously talked so much about humans in the loop. And in this latest piece, I noticed we seem to be involving in the direction of humans above the loop. What does that mean? Is that a relevant difference? Yes, it's a great distinction. I think to be clear, we will see both in the future. Human in the loop would suggest that agents or that AI is doing pieces of the puzzle in the process, then passes it to a human that does other pieces that they may pass it back or take it to completion. I think the human above the loop idea is what if we can get to a place where squads of agents or teams of agents are able to do most if not the entire core process. And the role of the humans is the judgment overlay on top. And let me give you an example. We've done reimagining of the legal arbitration process with the American Arbitration Association, when people send cases in for review where there's disputes about what the right answer should be. The traditional process for this involves gathering hundreds of not thousands of different data points. Sometimes these are photographic exhibits. They include contracts, often email exchanges back and forth. And then reviewing a case file and saying, what do I think the right answer is from a judgment standpoint, legally, based on the terms of the agreement? This can take a really long time. The question was, could we use an agentic reimagined process to really transform that and actually train agents using closed case files, so past cases, to put together the timeline, review all the fact-based look at both sides of the argument and come to a summary decision. What we found actually was that not only could these agents do a lot of that core work, but in some cases they could do it better. You still need a human, though, to look at that whole process and say, do I agree with the decision you've come to? It's still incredibly critical to have that, judgment layer, that human layer, but a lot of the work underneath can be done end-to-end by agentic AI. As the shape of the work begins to change, one of the harder questions is what that means for the people entering the workforce and for how judgment, skills, and experience get built over time. What happens when new talent is entering the system? How do these new employees, more junior in their tenure, develop the necessary skills to oversee AI without having gone through that hazing work process that the rest of us went through in the pre-agentic era? I think this is truly the billion-dollar question for companies. You eliminate every new software engineer. Not only do you have a very expensive model of only senior folks, but you roll that movie forward 10 years and you're missing the next generation that you need. The investment in learning and development is a bit of a sidecar. It's viewed as important. It's certainly valued by employees, but it sort of runs as this periodic thing. In the future, I think this is going to have to be at the center of the journey that people go on as an employee because the powerful thing is the next generation is going to have access to all these tools from day one. They may not have the pattern recognition of 20 years of looking at case law before them. They also don't have the hurdle of being 20 years into a career and trying to figure out a massively disruptive and powerful technology for the first time. The question is, how do I make that access from day one actually work to their advantage? I think the other aspect of it is really like capability building to help people see that change management is no longer going to be an episodic thing. It's going to be a perpetual state. We're going to have to get really good at being comfortable in constant change without introducing chaos and risk into the organization. Do you have a point of view on how these changes will affect the shape of an organization's talent, the structure of talent, not necessarily the traditional pyramid shape, but perhaps a diamond shape or a more obelisk shape? I think it's too early to have an answer on the final state of the shape of organizations. I think AI will absolutely enable that. But the team structure looks like there's a lot of excitement about this idea of organizing around pods of work as opposed to traditional job hierarchies. You can have these pods that form and reform and are reusable and move across the organization more nimbly, practically for companies. That's really hard. You need a job hierarchy because you need to evaluate people because they need to know who they go to for managerial support. I think we're still pretty far away from seeing most companies try to reorganize themselves around that principle. I do think this idea of fluid talent flows is going to increase. Organizations that have an HR function that enables that and can support really a more nimble movement of talent across parts of the organization that will become a competitive advantage. What are the implications for organizational culture in the context of transformation of this magnitude? In this moment, I think organizations that can really foster a sense of curiosity and continuous learning will be in a great space because so much of what is coming is still in the land of unknown. We need people who want to experiment and learn, people who can do joint problem solving, who can bring together people from across the organization, coach them on how to work into new models to explore the change with optimism and not fear. That I think will be a premium as well. Then that judgment that strikes the really critical balance between taking some risks and not betting the company because this is going to be one of those moments where we have to be highly iterative. We need two-way doors, not one-way doors. That's how I think about it. We don't have enough clarity to be walking through through too many irreversible one-way doors. We need situations where we can experiment and explore and then if and where it doesn't work, pull back, come back out and go into the next pathway. And beyond questions of structure leadership and talent, the value of any transformation like this has to show up in the work itself, in the workflows and processes where that value is created. Alexis, talk to us about what the implications for extracting value from an agentive transformation will be on workflows. We talk a lot about reimagining processes, reinventing workflows from end to end. What does this look like in practice? What we're seeing is at the best use cases where AI is really enabling scalable impact, is where a whole end to end workflow can be reimagined in its entirety. And why is that? It's because typically those workflows cut across multiple teams and multiple areas of a company. So inherently in a traditional context, they have a lot of touch points by a lot of people, a lot of rework and connecting the dots across. This is actually a place where agentic can be incredibly valuable because it can be part of that connection stream in a really fast and multifaceted way. And so for companies, what that means is rather than point solutions, where I think a lot of AI experimentation started it, let me take this one task and figure out how to do it with agentic or generative AI better faster and with higher fidelity. We're now talking about how do I take a whole workflow? How do I take something like insurance underwriting and actually rethink that end to end? How do I take the HR higher to onboard process and re-imagine that end to end? And what we're seeing is that's where the magic starts to happen. But that work is deep granular work. You have to rethink that whole process and say, where does it benefit from having an agent? Where does that need a human in the loop or above the loop to supervise? Where do I need a team of agents? How do I make those reusable? So once I train it, I can actually deploy it in multiple places. And that thinking is the system thinking that is really starting to unlock opportunities at scale. And I think that's the real exciting potential. But that requires a set of individuals who can do that type of problem solving and leaders who can pick areas that are really ripe to be those early lighthouse use cases. Just re-imagine processes in that way require folks from different levels of the hierarchy because unless you're really close to the process, you won't know whether the process will actually yield the result in the way that is usable. But what the composition of a team looked like, who is thinking about re-imagining certain processes? I think this is one of the really exciting opportunities with AI is this chance to bring together multiple levels of the workforce simultaneously. In this moment with AI, the potential is actually up and down the chain of command, if you will, or the hierarchy of the organization, which means you need people at all of those levels engaged in thinking about how that process could look different. Because a huge part of the unlock with a genetic in particular is actually that executive function, the review of results of work done, the governance on top. Now you still need human judgment overlaid on that, but you have the potential to take tasks that very, very often historically were done by managers or even senior leaders in companies and do much of that quality control review assessment for pattern recognition through the agentic process, which means you need all of those leaders at those levels down to the folks who are deep in the day-to-day flow of work to be involved in thinking about how that would look different. Alexis Cryptcovic, thanks so much for joining us today. It was such a pleasure. Thanks so much for listening to the McKinsey podcast. I'm Lucia Riley, and I'm Roberta Fissaro. Find us on McKinsey.com. We'll have a transcript of this episode up shortly. And download the McKinsey Insights app where you can find this podcast in other helpful content updated daily. If you enjoyed the show, we'd love for you to leave a rating and a review. We'll see you in two weeks. (upbeat music)