What the labor market isn’t telling you—yet

AI is fast transforming work as we’ve known it—and the latest data on jobs doesn’t always reflect the changes underway. “One word sums it up best: ‘uncertainty,’” says Svenja Gudell, chief economist of global employment platform Indeed. In this episode of McKinsey Talks Talent, Svenja joins McKinsey talent leaders Brooke Weddle and Bryan Hancock, along with Global Editorial Director Lucia Rahilly, to help leaders make sense of the current collision of labor market trends: generative AI, agentic AI, an aging workforce, shifting priorities, and more.

The following transcript has been edited for clarity and length.

Uncertainty in today’s labor market

Lucia Rahilly: Svenja, welcome to McKinsey Talks Talent.

Svenja Gudell: Thanks so much for having me. It’s a pleasure to be here.

Lucia Rahilly: Let’s start with a quick overview of what your role as chief economist of Indeed’s Hiring Lab entails.

Svenja Gudell: As chief economist for Indeed, I run the Indeed Hiring Lab, which is Indeed’s economics and research arm. Our mission is to be at the forefront of everything to do with the labor market. If there’s a particular trend happening, we want to know about it and to inform our stakeholders and audiences.

Lucia Rahilly: There is a lot of uncertainty in the macroeconomic environment right now. What are the most surprising trends you see emerging in the data for 2025, and how do they compare with what you’ve seen in previous years?

Svenja Gudell: One word sums it up best: “uncertainty.” The level of uncertainty we’re dealing with in today’s economic market is incredibly high. That’s because a few trends are colliding at the same time: politics and where economic policy is going, leftover impacts from the pandemic, and demographic shifts that are larger and more long term. Many industrialized countries have aging economies and shrinking labor forces, particularly in places like Japan, where this is happening acutely already. Then there’s AI, particularly generative AI, and how it is impacting all those trends.

Brooke Weddle: Svenja, as an economist, you look at a lot of data. How do you cut through the noise and say, “These are the two to three metrics that give me a sense for how any given labor market, whether in the US or another region, is changing”?

Svenja Gudell: We have an amazing playground of data. We’re able to look not only at the demand side—job postings, what employers are looking for, what kinds of skills are important, and how jobs are changing—but at the supply side as well—what job seekers are doing, what their behaviors are, what they’re bringing to the table, how they’re upskilling or reskilling themselves, and what they are expecting of work nowadays. Looking at that intersection is really important.

Then you have the larger macroeconomic environment. Particularly in the US, consumption is a big driver of what has made our economy historically so strong. That’s a nice forward-looking indicator. Sentiment is one of the big things we look at. And then standard stuff like unemployment, layoffs—particularly in this environment, where I feel like many of us are sitting at the edge of our seats and thinking, “Oh, what’s going to happen next? How are we going to react to this?” We have to look at the data to see, “Is there anything that’s telling us what might be upcoming?” The number of layoffs is, right now, one of the things we’re watching closely. Immigration numbers and general migration numbers are super important.

From churn to return

Bryan Hancock: What have you seen about people who are marginally attached to the labor market—the ones who are first to exit, first to come back in when the job market either gets weaker or stronger?

Svenja Gudell: Let me give you a little background, which helps frame where we are today. Initially, we had, of course, the COVID-19 shock to the labor market, and a lot of people exited the labor market.

Then we got vaccines. Things started speeding up again. And a lot of people reentered the market. We saw a whole lot of “unretiring,” if you will. People had said, “Alright, I’m retired.” And then either their nest egg wasn’t going as far as they thought, or they said, “Wages are climbing quite a bit right now. There’s a lot of upside to reengaging in the labor market.” So we saw really strong labor force participation again. And it carried all groups. Women started participating again, despite having to juggle—and still having to juggle—a lot of home and childcare duties.

Now we’re entering a slightly different market—and oftentimes a very different market, with return-to-work mandates and less flexibility. And as flexibility is removed, we’re not seeing the same wage increases. Now we’re in a steadier phase, and we’re not seeing as many people flow back into the labor market.

AI and jobs—from generative to agentic

Lucia Rahilly: Svenja, let’s get a sense of the holistic picture of the labor market that your data helps inform, starting with AI. Talk about how you see AI changing the game for job seekers. Are you seeing signs of displacement? Is AI reshaping the demand you see on your platform for certain kinds of skills?

Svenja Gudell: I’ll start on the macroeconomic side. The Hiring Lab team has done some extensive research trying to answer what I think is the number-one question out there: “How is generative AI impacting the labor market?”

We saw, particularly in the early days—three years ago—a lot of surveys that asked, “How will this impact different jobs? How are different skills and tasks going to be impacted?” We really wanted to take the wealth of data that we have at Indeed to bring a more quantitative approach to that question. So we dug in and tried to see, “Given all the skills in the Indeed skills taxonomy—close to 3,000 that basically describe the skills every single job needs—how are those skills impacted by generative AI?”

We looked not only at the general knowledge generative AI will bring to a certain skill but also at problem-solving abilities, as well as the need to be physically present. For example, as a nurse, if you’re taking blood, you have to be there to perform that action. So generative AI, at least currently, is not very good at doing something like that—although ten years from now, I fully expect a robotic nurse to take blood.

Bryan Hancock: How is generative AI changing the work of economists? How are you using gen AI in your work in looking at labor markets?

Svenja Gudell: It’s the next thing. Taking all that data, we were able to see that generative AI does not replace any job completely. There wasn’t a single job where generative AI was particularly good at all of it. However, it’s particularly good at doing a few things really well—for example, coding. Employers are adjusting how many software developers they hire or what kinds of efficiencies they’re starting to see in terms of how much code is written by generative AI.

The interesting outcome of that research was that we’re seeing a lot of augmentation of jobs. And the traditional list of impacted jobs was suddenly turned on its head. Automation impacted a lot of manual jobs. Generative AI impacted a lot of knowledge worker jobs: marketers, HR professionals, accountants, software developers. If you’re a bus driver or childcare professional or dentist, you saw less impact. If we have this conversation three years from now, I bet the landscape for jobs will have changed even more. And then, of course, there’s the other side: How is hiring changing, and what are the tools available to employers and job seekers as they go on their journey of finding that perfect job match or perfect candidate?

The usual thing for me to say for job security is that economists will never be extinct. I’m happy to report that so far, most generative AI tools aren’t really good at forecasting, or doing some of what we do daily, or even analyzing different trends and pulling them together. What it is really good at is summarizing research, for example. It’s faster now for me to do a literature review or get a sense of what the landscape is—what existing research is out there and where we might want to go.

Brooke Weddle: Svenja, what about agentic? Are you seeing any early signs of AI displacement, particularly among entry-level candidates entering the job market?

Svenja Gudell: Absolutely. We are seeing fewer internships, fewer entry-level jobs. We are also starting to see less demand for the highly impacted job functions, such as software developers. Interestingly, if you dig a little deeper, you’ll see a fair bit of demand for, say, data engineers—people who work with the data. You still need to feed your large language model a lot of data. But for pure coders, especially entry level—less so on the architect side of things—if you’re impacting a third of all that workload for every single person, you can probably combine those jobs. Maybe you hire one less person, and then another person does the more human aspects of the job that the machine cannot do yet. It’s a rejiggering of skills for a single job. But that impacts the type of job that is then hired for.

Bryan Hancock:  Given the challenge of reduced postings and reduced openings for entry-level workers, what’s happening to the unemployment rate for recent college graduates?

Svenja Gudell: Unemployment is creeping up a bit. And recent graduates are more impacted in some countries than in others. Some of that is driven not only by AI but also by macroeconomic trends. But it’s a good thing to pay attention to because it brings up the larger questions, “What will happen to the entry-level job as we incorporate more of these agentic experiences or more generative AI? Will we continue to see unemployment creeping up more than it currently is?”

The tricky part is, you can’t look at this in isolation. From the employer perspective, you’re thinking, “I have other tools to do a lot of these tasks that I used to have an entry-level hire do.” But from the perspective of someone looking for a job, there’s a lot of potential for uptraining or reskilling using these tools.

I think this is the first time we’ve seen a disruption be not only the disruptor but also potentially the vehicle to help you live through that disruption. And I expect that will feed into some of the capabilities job seekers bring to their applications and to jobs later on. We have a ton of profile data at Indeed—over four million—that we look at from a research perspective. And one of the things we’ve started to pick up on is that a lot of folks working at McDonald’s are listing generative AI skills on their résumé. So I think we’ll start to notice people are getting in there and saying, “How can I use this tool to help me perform better in the labor market?”

Lucia Rahilly: Svenja, we’ve been talking about generative AI obviating certain tasks or reducing and reconfiguring roles. Any hot new roles emerging as AI starts to advance toward greater scale?

Svenja Gudell: Certainly. And history is on our side here; we’ve seen this story before. Every time there is a big technological development, we see the destruction of jobs and then the creation of new ones.

We’ve all talked about the prompt engineer, which was a big thing—it’s not that new anymore. We’ve seen a pickup in AI consultants. We’ve seen more salespeople having to be trained on selling AI technology. So some of these worlds are colliding and new jobs are being created.

From an employer perspective, this can often be tricky. Employers think, “I have to find people who understand AI and can also sell a product. Where do I find those skills together?” That brings us back to that original point around uncertainty.

If you look at payroll data and the jobs report, a lot of new jobs have been part-time. We’re seeing employers reach for a tool that they have in times of uncertainty. Maybe they don’t want to increase by a full headcount, but they’ll hire a contractor or a part-time employee to keep their options open and really flex in this environment of uncertainty.

McKinsey Talks Talent Podcast

Bryan Hancock, Brooke Weddle, and other talent experts help you navigate a fast-changing landscape and prepare for the future of work by making talent a competitive advantage.

See more episodes

Evolving priorities

Brooke Weddle: Svenja, I want to return to your interesting insight around more McDonald’s workers sharing that they have a skill set that includes generative AI. How have you seen the role of the résumé change over time? What should job seekers be emphasizing, or what should they be thinking about differently when it comes to that core tool of getting a job?

Svenja Gudell: One of the really big trends is skills first. The résumé is still a handy tool, and it’s still very widely used. Does your résumé need to live on a piece of paper? Maybe not. Maybe it’s just a collection of things you enter on a website, so it’s faster for a recruiter to gain appreciation for everything a candidate brings to the table: “Can you actually do what is required of you? Do you bring the right skills?” versus “Where did you go to school? How many years of experience do you have?”

That shows up in our data by how many jobs still require a master’s degree or a BA, for example, or how many jobs specify “ten-plus years’ experience” or “two years’ experience.” That number is going down—not as rapidly as in a very competitive market, because it was also a way to expand your talent pool. But we’re still seeing that push toward skills first.

Lucia Rahilly: I’m wondering about return to office—a hot headline at the outset of 2025. Is talent still prioritizing hybrid and other kinds of flexibility? Any changes in what job seekers are optimizing for?

Svenja Gudell: If you’re looking at the preferences and needs that a lot of job seekers have, number one is always compensation. Everyone wants a better pay package. No surprise there.

But particularly for women, the number-two reason for switching jobs remains flexibility—including either remote or hybrid working arrangements. We’re still seeing that as really important to a lot of job seekers. However, the number of jobs that offer remote work in the US has been steadily declining—perhaps not as much as the headlines might suggest, but from a high of somewhere around 10 percent to, I think, below 8 percent.

Some of that has to do with job mix. Right now, you have fewer openings in tech-type jobs, which tend to have a remote angle. We’ve seen a decrease in availability of remote and hybrid jobs on the supply side but not in terms of demand from job seekers.

Lucia Rahilly: I want to pick up on something you said about the aging population. How does generation factor into the way workers are engaging with the labor market? Everyone’s focused on Gen Z. Does the data show that Gen Z has demonstrably different priorities in the jobs they’re pursuing?

Svenja Gudell: I think generational differences are not necessarily about the types of jobs candidates are searching for but rather what they’re hoping to find on the job. For example, we’ve done quite a bit of research on how political issues creep into life at work. Employees have expected companies to take more of a stance on certain issues. But there can be a lot of different viewpoints within the company itself, so the research is really interesting.

We also ran a survey that looked at how different generations look at diversity, equity, and inclusion—how they think about diversity among their senior leaders, how they think about bringing their whole selves to work. And there’s definitely a trend where younger generations care about this more than older generations. Part of that could be just what you’re used to and the trends that have been predominant at different times in the labor market.

What leaders are missing—and why it matters

Brooke Weddle: This gets me thinking. We’ve talked about a couple of headlines: return to office, AI. Are you looking at any trends and saying, “Gosh, why aren’t people talking about this or spending more time on this from a thought leadership perspective?”

Svenja Gudell: A lot of people really tend to focus on the short term—the now—and not really on having a good, long-term approach. And to survive, you have to have a short-term view on some things. You want to be able to open up your doors tomorrow as an employer. But having long-term trends in mind is also really important. So that’s always surprising to me, how headline driven we are.

And then as an economist, one of the really surprising things to me is, every time new data releases come out, we ask, “Oh, is this going to be the one that shows the impact, for example, of tariffs?” A lot of these trends take months to show up in official data, which makes us reach for other data sources. And then people are like, “Oh, we’re good—no inflation impact yet” or, “Consumption looks just fine.” But that data is three months old. It won’t show the impact.

With the onset of AI, one of the things we’ve looked at quite a bit is the “technification” of jobs. We have an AI tracker that looks at how many job postings mention AI in the job description. That could be for a creator of AI, someone who codes it, or a user of AI—someone in marketing or HR who uses that technology. And our tracker still has less than 1 percent of all jobs on Indeed that mention any sort of generative AI keywords. So it’s a really low number but really strong growth. Then you look at the technical skills most jobs require. And surprise, surprise: Microsoft Office is still the number-one-mentioned technological tool that every job wants out there. It’s really surprising.

Lucia Rahilly: Svenja, you mentioned seeing some slowing in the labor market, particularly in certain segments like entry-level jobs, as well as slowing attrition due to uncertainty. Acknowledging those dynamics, the labor market is still relatively tight. What are the biggest challenges you see for employers looking to attract and retain good talent in today’s marketplace?

Svenja Gudell: To me, the number-one thing to recognize is that the labor market is not a giant monolith. There are some really interesting things, particularly in the tighter market, that you can look at. Often, you might not have flexibility on base pay, but you have flexibility on benefits. We’re still seeing signing bonuses in the healthcare sector quite a bit. You can offer different training programs that help attract talent. Remote work is still a big thing that carries weight. A lot of different tools that an employer has in their toolbox can help in a tight market—beyond just paying more money.

The labor market of tomorrow

Lucia Rahilly: As you’re looking out beyond 2025, do you see any nascent trends or disruptions that business leaders should be aware of or act on now to get ahead of?

Svenja Gudell: Absolutely. The big one, again, is the race against an aging workforce. We’re going to start to see people still demanding goods and services without being active in the labor force anymore. So the number of workers actually available to take jobs is going to start to come down.

Balance that with AI. Are we going to see some of these promised productivity increases along the way? Which path will win out? Will we get AI up and running? It’s easy to have the technology, but you have to be able to use it productively inside your firm. Will that happen before we start to see these impacts of an aging workforce?

Those are the big ones. And then digging deeper there, I think every employer should—and many do—ask the questions, “How should I hire differently today? I hired a person five years ago, and this was their profile. Should that profile look different? Should I not even hire this person anymore, because these skills won’t be needed or won’t need to be performed by a human a year from now?”

That workforce-planning aspect, I think, is super important. And for that, you’ve really got to have more than your big toe dipped into the AI pool to understand what’s really happening.

Brooke Weddle: That really resonates. As you think about trends and the adoption of AI, how will that impact employee sentiment?

Svenja Gudell: I’ve had the really cool opportunity to go to Davos for the past two years. In 2024, generative AI was the talk of the town. Everyone was kind of fearful, on the side of both the employer—“How can I integrate this and make the most of it?”—and the employee—“Is my job going to go away?” This year had much more of a flavor of, particularly for agentic AI, “This will be a partner for you. You work alongside the tool to produce things.” And some of that is driven by the fact that generative AI still lacks some very important human traits like empathy, leadership, parts of communication.

So how do you marry those two? That is a really interesting path to go down: “Have employees taken that opportunity to be able to use the tool to their advantage?” We see this in the data; they’re slightly less afraid of it as they’ve learned more about it.

Lucia Rahilly: Thanks so much for joining us, Svenja.

Svenja Gudell: I very much enjoyed the conversation. Thank you.