Generative AI can give you “superpowers,” new McKinsey research finds

Meet the leadership team behind our flagship research, which looked at generative AI in terms of productivity gains for the economy, businesses, and individual workers.

Over the past six months, workers across industries, from marketers and developers to product designers, have been discovering the ways in which generative AI can help them do their jobs better. This could mean creating extraordinary content in a fraction of the time, accelerating IT coding and testing processes, optimizing product simulations and design for much higher quality. In short, generative AI can give them “superpowers.”

In our newest report, “The economic potential of generative AI: The next productivity frontier,” the Firm’s researchers also project that generative AI can add the equivalent of $2.6 trillion to $4.4 trillion to the economy annually.

“It's very compelling on a personal level because almost everyone can use it,” explains Michael Chui, a McKinsey Global Institute (MGI) Partner. “Generative AI has captured the imagination of business leaders in a way many tech trends don’t. If I talk about the cloud or quantum computing, it’s not something someone can touch. But with ChatGPT, you can open your browser and start asking it questions.”

With the right prompts, ChatGPT can write a short story, summarize 1000 pages into a five-point plan, or even compose a song.

While generative AI is off to a blazing start, we know from experience that gaining the full value from a technology and implementing it across an organization takes time, talent, and hard work.

Here’s what three of our experts— Alexander Sukharevsky, Senior Partner and global coleader of QuantumBlack, AI by McKinsey, Michael Chui, and Lareina Yee, a McKinsey Senior Partner, have learned from the research and from early adopters.

Let’s start with the big question: What is McKinsey doing about generative AI right now?

Alexander: We have been investing in AI for many years and generative AI is not new for us; our first investment in the space was around five years ago. Our focus remains on creating and scaling enterprise-grade solutions. To do this, we have a deep talent pool—more than 1,500 data scientists and engineers in more than 50 countries—and an extensive ecosystem of external partners. Earlier this month, we launched QuantumBlack Horizon, a suite of solutions specifically created to help businesses scale AI, including generative AI, in a secure, cost effective way.

From left, Lareina Yee, McKinsey Senior Partner; Michael Chui, MGI Partner; and Alexander Sukharevsky, McKinsey Senior Partner and global coleader of QuantumBlack, AI by McKinsey

What can we learn from early adopters’ experiences?

Alexander: The CEO conversation has changed from a few weeks ago from “What is this?” to “What do I do about it in my organization and how to start?” On the strategic level, first you need to bring your best talent together—a multidisciplinary group of technical and nontechnical people—to learn about the technology to establish a common language and to reimagine your business model and potentially redefine your respective industry. It doesn’t have to take long—I believe this is a real “age of creators.”

Multiple industries will be completely revamped by using AI to lower the barriers to entry, to disintermediate company structure, or to deliver the same value propositions at a fraction of the current cost. The way organizations define the opportunities and the execution speed will differentiate the future winners.

We’ve seen some fatigue around AI implementations, but the energy created by generative AI has brought AI projects back into the spotlight and reinforced the need for companies to fix the same basics—the quality and availability of data, the IT architecture, technology and translator capabilities, and rewired operating models. It's not easy, but essential for long-term success.

Let’s talk a bit about the research and key findings.

Michael: We used a methodology established in 2017 to analyze how automation technologies are impacting the labor force. We looked at 850 occupations across 47 countries, covering about 80 percent of the world's workforce. We then broke down each occupation into 20 or 30 “sub-activities.” In terms of generative AI, we analyzed tasks such as understanding natural language, communicating with others, searching for and gathering information.

We estimate that about 60-70 percent of the time people spend working has the theoretical potential to be transformed by a combination of generative AI with other technologies. While it could take years for this to be realized, the potential is huge. In fact, we updated our models for the rate at which these technologies might be deployed, and in some scenarios, generative AI could accelerate the pace of adoption by a decade, compared to our previous estimates.

In terms of use cases, while there is potential across the enterprise, 75 percent of the potential value from generative AI is concentrated in four functions: customer operations, marketing and sales, software engineering, and R&D.

What does this mean for the average worker?

Lareina: We say that generative AI gives people “superpowers.” What we mean is that it can automate many of the routine tasks that people do during their workday, so now they can be more productive and do more interesting work. The technology can function as your coding assistant, your researcher, even your editor.

In early use cases in call centers, we see generative AI helping representatives with less tenure, say at level one, advance to a level four much faster. With AI, they have techniques similar to those of their higher-skilled counterparts, such as quality scripts and detailed customer context, preparing them to handle increasingly complex situations more quickly. They spend less time creating materials, more time with their customers. In general, people are pretty excited about this.

If you are a software developer, generative AI can do the more routine tasks of converting legacy code, debugging and testing, so you can spend more time developing new functionality.

Wealth managers can spend more time advising their clients, rather than analyzing and summarizing large quantities of technical content.

In general, workers can spend more time on the human connection and interaction aspects of their roles, which is something AI can't take away.

What are the implications for organizations as a whole?

Michael: It means there's going to be a lot of change as we all will have to shift the activities of what we do. Some jobs will change, some will go away, and new ones will be created, such as a prompt engineer. It’s a huge reskilling challenge and opportunity.

On the other hand, it could provide a time surplus that may be used to create a different work-life balance—it could even be the start of the four-day work week.