EY Global Vice Chair: AI’s biggest paradox comes down to what AI can’t do

For two decades, Ernst & Young has operated under the principle that human judgment is at the core of its business. Clients rely on EY not only for data analysis, but also for its unique perspective and insights that help navigate uncertainty and make better decisions. This is precisely the type of skill that artificial intelligence is now demanding in every role, from analyst to machine operator.

The paradox of the AI era is that as the technology becomes more capable, the market rewards the human qualities it cannot replicate.

While companies are investing heavily in AI, they are overlooking up to 40% of potential productivity gains due to a lack of investment in the complementary human skills needed for effective implementation. Traditional training programs often focus on technical expertise, leaving out critical abilities that will determine the success of an AI strategy.

AI excels at repetitive tasks and data analysis, but the human element remains crucial. People are responsible for defining goals, setting purpose, and providing context that gives data meaning. Humans make decisions about which objectives to pursue, recognize when context changes, and remain accountable for outcomes. Without human direction, data and insights may not be utilized effectively, and their full value may not be realized.

To harness the full potential of AI, it's essential to integrate AI fluency as a core leadership skill. This involves not only operating AI tools proficiently, but also evaluating their outputs for relevance and knowing when to intervene. With machine identities outnumbering human employees by a ratio of 82:1 in most organizations, the responsibility for human oversight becomes even more pronounced. As AI systems become more autonomous, human judgment becomes increasingly valuable.

Analytical thinking has emerged as one of the most sought-after core skills for employers, with roughly 70% of employers considering it essential. Leaders now want employees who can avoid blind delegation and challenge AI when necessary. The bottom line is to use AI to improve efficiency, not to replace the judgment decisions that remain at the heart of the job.

Collaboration is key in navigating the AI workforce integration. Isolated approaches can be time-consuming and costly, as organizations waste resources rebuilding workstreams that have already been solved by others. EY's recent work with beverage manufacturer Lion demonstrated the benefits of collaboration, resulting in a 75% faster customer response time and a 30% improvement in operating costs across its people function.

By leveraging existing solutions from ecosystem partners, Lion was able to scale AI technology more efficiently.

To maintain a human-centered approach, workflows must be redesigned to keep humans at the core of decision-making. Rigid handoffs expecting AI to complete tasks before passing the results to a human can hinder meaningful judgment calls and add unnecessary iterative work. Instead, a more durable model involves organizing work around shared tasks with a truly integrated process where humans and AI contribute at every stage.

Daikin, a multinational heating, ventilation, and air conditioning company, successfully implemented this approach by collaborating with AI and human teams to roll out a new enterprise resource planning (ERP) system. While AI assisted with code generation and automated testing, employees maintained oversight for exceptions and high-risk scenarios, emphasizing that judgment is something algorithms cannot absorb. This approach accelerated delivery by 30% and improved counter efficiency by 10% and financial close time by 20%.

The skills that don't become obsolete with the evolution of AI are those tied to judgment and human oversight. As AI model capabilities constantly change, the conversation around upskilling should focus on the ability to know when and how to intervene, rather than simply acquiring technical updates. Organizations that successfully scale AI are making deliberate choices about which aspects of the job remain human. In the end, AI adoption done well clarifies the role of human judgment rather than diminishing it.