McKinsey Quarterly

The human side of generative AI: Creating a path to productivity

| Article

Ever since OpenAI’s ChatGPT exploded into public view in late 2022, the possibilities of generative AI (gen AI) have captured imaginations throughout the business world.

When it comes to crafting an effective talent strategy, organizations have focused mostly on how gen AI can increase productivity levels. This is understandable, given the trillions in value at stake. However, it may not be the most strategic approach. To match the right talent to jobs, leaders first must understand how gen AI is changing the way employees view their work experience.1

McKinsey recently surveyed a cross-section of employees as part of our continuing research into how organizations can improve workforce engagement, retention, and attraction (see sidebar, “About the research”). Respondents provided several intriguing insights that can help organizations as they build gen AI talent capabilities.

  • In any given organization, the pool of gen AI talent is likely broader than many leaders realize—and it’s poised to grow rapidly. This cohort isn’t limited to technical talent such as data scientists, software engineers, and machine learning specialists, important as those roles are. In fact, just 12 percent of our respondents fall into this tech-heavy category of traditional gen AI talent. The vast remainder of respondents, or 88 percent, are in nontechnical jobs that use gen AI for help with rote tasks. These jobs include middle managers, healthcare workers, educators, and administrators, among others (Exhibit 1).
  • Fifty-one percent of respondents in technical and nontechnical roles who identify as gen AI creators and heavy users of the technology say they plan to quit their jobs over the next three to six months. This is sobering news for those executive respondents in the survey who say they want to build gen AI talent in-house; it’s hard to reskill and upskill people when they are looking to leave.
  • Although those who self-identify as heavy users and creators of gen AI represent an in-demand employee group, these workers aren’t staying in jobs or attracted to them because of compensation. In fact, the survey shows that this group strongly emphasizes flexibility and relational factors such as meaningful work, caring leaders, and health and well-being over pay.
  • Finally, and perhaps most surprising: heavy users and creators of gen AI overwhelmingly feel they need higher-level cognitive and social-emotional skills2 to do their jobs, more than they need to build technological skills. As workers increasingly use gen AI to tackle more repetitive tasks, the human-centric skills of critical thinking and decision making will become ever more important.
Workers who use generative AI as part of their jobs comprise a much larger group than those who hold traditionally technical roles.

These revelations have broad implications for employers as they try to attract and engage their workforces. Organizations are on the cusp of gen AI pushing either positive or negative change when it comes to the nature of work. Leaders have an opportunity to humanize that work by deciding where, when, and how their teams use gen AI so that people are freed up from routine tasks to do more creative, collaborative, and innovative thinking. Gen AI talent agrees.

In this article, we break down crucial segments of workers who are at the forefront of gen AI usage or creation and dig deeper into the job factors and skills they say they need. We then discuss how organizations can enhance productivity by crafting jobs that put people before tech—not the other way around. Companies that set a people-centric talent strategy will give themselves a competitive edge as more workers and jobs are affected by the changes gen AI brings.

The workforce: Who is in the gen AI mix?

If companies are to take advantage of the productivity gains from gen AI, they first must consider the broad range of skills required for its successful deployment across the enterprise.

While there are many categories of workers who can be described as gen AI talent, we focus on four distinct archetypes in our survey based on gen AI use:

Creators: These employees help build the gen AI models for their organizations and develop the tools and interfaces most of us use to interact with these models. Creators (2 percent of employees surveyed) tend to be predominantly software engineers, programmers, and machine learning scientists who develop the tools and interfaces most of us use to interact with gen AI.

Heavy users: These employees use gen AI to help them perform most of their core tasks or to enhance their work functions. Heavy users (8 percent of our sample) include a wide range of workers, from designers who use gen AI to expedite 3D modeling to data scientists who use gen AI to verify the accuracy of their coding language semantics.

Light users: Workers in this category use gen AI to perform less than 50 percent of their primary tasks. Representing about 18 percent of the sample, they include middle managers, educators, and communications professionals. For example, a manager might use gen AI to create meeting notes or to help delegate tasks, while a teacher may use it to innovate classroom activities. Journalists and writers researching topics might use gen AI to give them a baseline of facts or to help write a first draft.

Nonusers: These are individuals who are either unaffected by or unaware of the impact of gen AI on their jobs. Examples in our sample include nurse practitioners and healthcare workers engaged in direct patient care, as well as retail associates whose primary role is face-to-face interactions with customers. Although these employees currently represent about 70 percent of the survey, our expectation is that a majority of nonusers will become light or heavy users as the scope and usage of gen AI changes.

Never just tech

Creating value beyond the hype

Let’s deliver on the promise of technology from strategy to scale.

People over pay: The job factors that workers value most

The COVID-19 pandemic revealed that for many workers, what they want most from their work experience has fundamentally changed. Employees increasingly value relational elements such as caring leaders and coworkers, as well as support for health and well-being, more than compensation (though pay is always important). In 2021, we saw workers quitting in droves—in fact, 40 percent of respondents across jobs, industries, and geographies said they planned to quit their jobs in the next three to six months. That figure has since dropped to 34 percent.

Certain worker segments, however, remain a greater flight risk. Of self-identified gen AI creators and heavy users, 51 percent of respondents to our latest survey say they plan to leave in the next three to six months.

Early creators and heavy adopters, in particular, wield power when it comes to job choice and shaping their careers. Many company leaders believe that workers in these groups are leaving at higher rates because they can find better compensation elsewhere. Yet an examination of the employee-value-proposition (EVP) factors that resonate most with these segments busts the myth, once again, that compensation is a primary motivator.

Our survey shows that creators and heavy users prioritize workplace flexibility more than total compensation, and that they are seeking a sense of belonging, care, and reliability within their work community. They stay in their jobs when they are given flexibility, and they leave when they aren’t. The other factors that make them stay are meaningful work, support for health and well-being, reliable and supportive coworkers, and a safe workplace environment. This experience is similar to what most workers want, with one glaring exception: compensation appears much further down the list (Exhibit 2).

Creators and heavy users of generative AI who plan to stay in their jobs prioritize flexibility and relational factors over compensation.

McKinsey analysis shows that high disengagement and dissatisfaction rates can cost companies millions of dollars a year. Broadly speaking, addressing why workers stay or go is therefore paramount for companies as the use of gen AI grows.

When we dig deeper into self-identified heavy users and creators who are staying in their jobs, we find that a healthy 72 percent report feeling engaged at work, compared with 63 percent in our total survey sample. However, a worrying 55 percent report clinical levels of burnout, a much higher rate than the global sample of 32 percent. In other words, companies may not be getting the productivity and engagement they expect from these workers.

These EVP elements also play a big part in steering workers into new positions. For the broader workforce, the top four factors for why people take a job are similar to why they stay. However, for workers who identify as heavy users and creators of gen AI, there is a stronger emphasis on relationships with managers and peers, and on a sense of community more broadly.

Specifically, half say that reliable and supportive people are crucial, and nearly half emphasize the importance of caring and inspiring leaders. Roughly two in five say that meaningful work and an inclusive community are core motivators, even above flexibility, which registered as of primary importance to those staying in their jobs. In contrast to the broader set of workers where compensation is the third most important attractor, for this subpopulation it again ranks seventh as a motivating factor. People won’t come just for the money, and they certainly won’t stay for it (Exhibit 3).

Jobs that feature reliable and supportive people, caring leaders, and meaningful work are attractive to creators and heavy users of generative AI.

Most wanted: Cognitive and social-emotional skills

As gen AI interaction deepens (moving from nonuse to light use to heavy use), we see a consistent trend among both technical and nontechnical workers: they rate higher cognitive skills as more important than technological skills. Even among the technical workers who identify as gen AI creators, higher cognitive skills, at 59 percent, are rated as more important than technological skills, at 55 percent (Exhibit 4).

Cognitive and social-emotional skills are more important than technological skills across a number of categories for users and nonusers of generative AI.

Regarding social-emotional skills, two interesting trends emerge. First, most technical talent sees social-emotional skills rise in importance as this group increases its usage of gen AI, while nontechnical talent reports the opposite trend. Second, creators who identify as technical talent report lower importance for social-emotional skills at a similar level to nonusers.

Taken together, it appears that as workers become more heavily involved with gen AI, their focus shifts away from social-emotional skills, unless they are in technical positions. It may be that workers are unaware of how their jobs will change in relation to managing and interacting with other people, particularly regarding the importance of developing crucial social-emotional skills.

The disconnect: Employers want to build gen AI talent mostly in-house

Many companies are striving for the most effective way to solve the supply–demand issue when it comes to gen AI talent. Our survey of executives found that most organizations plan to build their gen AI capabilities internally, through upskilling, reskilling, and redeploying talent, more than by external hiring and contracting (Exhibit 5). Naturally, given the spread of worker archetypes in organizations and the workforce more broadly, some subpopulations, such as programmers and software engineers, may be best brought in through hiring while other types of workers, such as associates and customer experience specialists, will benefit more from upskilling and reskilling to bridge the gap.

More employers expect to close generative AI skills gaps internally.

The problem is that if companies want to build gen AI skills with the employees they already have, they need to retain the very people who, according to the survey, have indicated that they plan to leave in the next three to six months.

This gap between what employees say they want in a job and what employers are willing to offer them has been central to the workplace experience since the pandemic erupted. Our talent trends research has found that employees consistently want flexibility and meaningful work, and they want to feel valued and engaged.

When mapping self-identified heavy users and creators of gen AI onto which EVP factors matter most, we see that their emphasis on relational factors is largely the same as our broader survey sample. The need to care for family shows the largest increase in importance, while compensation registers the largest decrease.

Additionally, feeling valued by a manager, having access to development opportunities, and doing meaningful work also show a notable increase in importance. Advancement opportunities, on the other hand, are not as highly valued, suggesting that there are some unique conditions to being in a highly technical job, either through the creation of gen AI or through its heavy use (Exhibit 6).

There is little doubt that gen AI can help increase individual and workforce productivity; it may automate up to 30 percent of business activities across occupations by 2030.

How leaders can close the gap

There is little doubt that gen AI can help increase individual and workforce productivity; McKinsey research suggests it may well automate up to 30 percent of business activities across occupations by 2030.

Leaders should explore answers to three fundamental questions about their workforces in light of the impact of gen AI:

How can we reimagine jobs to be more human centric? Begin by defining which tasks people should do, which tasks gen AI can do, and how humans should manage other people as well as gen AI usage itself. Technological skills such as coding will be the baseline for many jobs, but social-emotional skills and higher cognitive skills will be the differentiators for creative, collaborative work in the future. Perhaps this means more in-office meetings or other ways for people to engage in the most productive ways they can.

Workers who perform at high levels and inspire others—we call them “thriving stars”—help spur collaboration, innovation, and better decision making. However, they make up as little as 4 percent of organizations. Their scarcity makes it particularly important to place these employees in positions that will boost overall performance.

How can we redefine flexibility? As jobs change, companies will need to look at worker outcomes according to the results achieved, not by hours spent. The benchmark for output will have to shift. For instance, some written code may be longer, but it may not necessarily be better or more user friendly.

With the potential for gen AI to help make jobs more efficient, could an employee’s meaningful work in a given week be completed in as little as 20 hours? And if that’s the case, is the 40-hour workweek still the benchmark? Rather than filling hours with tasks to get to a specific number in a given week, companies can focus on ways to emphasize the distinctive, creative part of a job that makes it meaningful. Jobs that create the space for the human touch can also help facilitate a more engaged and more productive workforce.

How do we emphasize the right kind of listening? This is a basic concept that many organizations seem to have trouble embracing: talking with employees rather than leading by assumption. Creating a constantly evolving dialogue can help with both problem solving and morale. This is particularly relevant as the gen AI talent pool expands.

Survey respondents overwhelmingly express enthusiasm about the integration of gen AI into their workplaces, though approximately 4 percent say they are concerned about job displacement (rising to 7 percent for workers aged 18 to 24). This undercurrent of worry presents an opportunity for leaders to engage workers about the potential changes gen AI will bring.

To illustrate how these shifts apply to the workforce today, we offer two examples of nontechnical gen AI talent: a communications specialist and a middle manager.

More time for innovation and collaboration

A communications specialist in a large corporation is currently a heavy user of gen AI. Her job has involved interviewing C-suite executives and synthesizing their ideas to create speeches, talking points, emails, and other communications for both internal and external audiences. Her performance has been measured by how many discrete communications she facilitates and the quality of the copy that is produced.

She used to send questions to executives ahead of time and then schedule a series of interviews, which would take several weeks to complete. Now, she can feed their recorded interviews into a gen AI chatbot and get a synthesis of their remarks in seconds.

The communications expert will still review and edit that text, but the overall process is much faster. Whereas before she spent 60 percent of her time synthesizing material, that task now takes only 10 percent of her time, freeing up bandwidth to think strategically about the message the speech is intended to convey and what form of communication would be most effective. She may also have more time to deepen relationships with industry reporters, which could benefit coverage of the company, and to help the chief human resources officer write that book she has been eager to start.

This gen-AI-related efficiency gain leads to increased productivity, more innovative thinking, and welcome face time with key constituencies—good for the employee, her team, and the organization. The value she adds to the job is now fundamentally different.

Managing people, managing gen AI

Now, consider a middle manager at a technology company who identifies as a nontechnical creator of gen AI. Currently, middle managers report spending almost half of their time on individual-contributor and administrative tasks and only about a quarter of their time on people-related activities. In a gen-AI-enabled world, they could significantly reduce the number of hours spent on non-people-related activities and reallocate that time toward supporting direct reports and engaging in broader strategy concerns.

As teams start using gen AI to help free up their capacity, the middle manager’s job will evolve to managing both people and the use of this technology to enhance their output. In other words, gen AI will become another member of the team to be managed. And just like a direct report who needs some intensive coaching to get up to speed, gen AI may need more guidance and involvement from managers—at least initially and perhaps for much longer.

Lastly, a core part of the manager’s role will be to ensure the humanization of work. As the nature of tasks and time spent change, and the focus shifts from process oriented to results oriented, managers will be a decisive factor in whether an organization allows gen AI to elevate people’s work. Keeping a finger on the pulse of their teams raises the likelihood that managers will do their part to create jobs that are less abstract and disconnected and more fulfilling and collaborative. To prepare people, managers can encourage employees to recognize the centrality of their insights and creative contributions with respect to the broader organization as gen AI use evolves.

The employer–employee disconnect has led to high levels of workforce discontent, which is affecting workers at the forefront of the gen AI push even more dramatically when it comes to burnout and attrition. Companies that want to capitalize on gen-AI-fueled productivity gains have an opportunity to address this rapidly expanding group’s concerns about the nature of work. Those that emphasize the importance of human skills over a simple race for increased output are likely to earn the loyalty of their workforces and higher performance over the long term.

Explore a career with us