Major new technologies have always changed the way people work, but artificial intelligence could be the most powerful yet. Many fear this will mean massive job losses. Our research suggests something more subtle, and ultimately more hopeful. In its current form, AI works best in collaboration with people, rather than on their behalf. To make that collaboration successful and ensure people thrive, organizations need to rethink how they develop and leverage skills, from frontline employees to the C-suite.
A new study from the McKinsey Global Institute (MGI) finds that currently available AI-powered agents and robots are technically capable of performing tasks that take up about half of current U.S. work hours. This is a surprising number, but one that is easily misunderstood. This does not mean that half of all jobs will disappear; it does mean that many of today’s work activities will be automated, allowing people to perform different parts of the same job, and potentially creating new jobs. How quickly these changes occur will depend on business choices regarding technology adoption. If the history of other technological transformations is any guide, it could take decades to be fully implemented.
Still, the possibilities are huge. We estimate that the combination of humans and AI tools such as agents and robots could unlock approximately $3 trillion in economic value in the United States by 2030. Realizing these benefits requires rethinking work itself. This means rethinking processes, roles, and culture so that humans and intelligent machines can work together to significantly improve productivity.
Success will require not only the technical capabilities to build and manage AI, but also new skills, including the human and analytical capabilities needed to work effectively with these tools. Over the past two years, job postings have shown a sevenfold increase in demand for AI fluency, or the ability to use and manage AI. This skill is increasingly in demand in a variety of roles, including SEO specialists, organic chemists, financial reporting managers, and engineers, an early sign of broader changes to come.
In the short term, skills such as conflict resolution and design thinking will remain as valuable as they are today, as they rely on creativity, situational understanding, and empathy. These skills cannot yet be replicated at a high level by machines. Other processes, such as accounting and insurance claims reconciliation, may be better handled by AI. Other tasks, such as research, data analysis, and project management, can be performed by humans or machines, depending on the situation.
MGI’s analysis found that approximately 72% of today’s skills can be applied to tasks that can be performed by both humans and automation. In other words, most existing skills will remain relevant, but people will use them differently as machines take on some of the tasks that those skills once supported. This includes many foundational skills such as communication, management, operations, problem-solving, leadership, attention to detail, customer relations, and writing that will persist and evolve as humans and AI collaborate.
The difficult task for organizations is to carefully combine people and AI to maximize each other’s strengths. This means building AI fluency across the workforce, from frontline employees to senior leaders, and designing complementary roles for machines to handle daily tasks. At the same time, people set goals, guide them and check their achievements.
Think about problem solving. Identify problems and find solutions. Agents and robots can drill down into datasets to identify patterns and generate options. But it’s up to people to interpret those findings and decide what to do. For example, in customer relations, agents and robots can handle routine queries, freeing people to focus on building stronger relationships. In healthcare and life sciences, generative AI is beginning to help medical writers edit and draft reports, while people review output, apply clinical judgment, and ensure accuracy.
The future will be shaped by how smart humans and intelligent machines work together. That would be different from the way people work today. Getting there starts by imagining the future and working backwards to redesign your entire workflow. Incremental change alone is not enough.
After analyzing 190 business processes, we found that approximately 60% of the potential productivity gains are concentrated in sector-specific capabilities, or the core activities of each industry. In manufacturing, opportunities center around supply chain management. Healthcare, clinical diagnosis, patient care. and in finance, regulatory compliance and risk management. The remaining profits come from cross-cutting functions that support operations across the industry, including IT, finance, and managed services.
Major companies are already experimenting. A global technology company uses AI to redesign its sales workflow from start to finish. First, our prioritization agent uses extensive public and proprietary data to score and rank your accounts. Other agents then help with customer interactions, and scheduling agents set up calls and reminders for potential leads. Sales reps then manage customer discussions and orchestrate the entire portfolio of opportunities. The result is a 7-12% increase in revenue and approximately 40-50% time savings that can be reinvested in more direct service to clients. Of course, not all efforts are this simple, and the company is still learning how to rethink its processes, but this example provides a glimpse of what’s possible.
AI has the potential to transform business and the way we work. Successful leaders leverage technology to enhance human potential. Retrain people for higher-value jobs, organize around skills rather than job titles, and measure success by how effectively people and machines work together to produce better results.
The opinions expressed in Fortune.com commentary articles are solely those of the author and do not necessarily reflect the author’s opinions or beliefs. luck.
