AI has not yet created the productivity boost as expected: the road to work redesign is still very long

• More than 3 years since the inception of ChatGPT, many businesses have yet to see AI create the clear leap in labor productivity they expected.
• At the Davos 2026 Forum, the topic “why AI hasn’t had an impact yet” appeared in almost every discussion among top leaders.
• A major reason is incomplete AI adoption by employees, due to fear of job loss or skepticism about the quality of AI tools.
• Many companies have forced AI usage through mandatory training or by tying it to performance reviews, but this approach has backfired.
• Cisco stated that mandatory AI training does not help sustain long-term usage and can even have negative effects.
• A more effective way is to give employees choice, such as providing a variety of AI tools and letting them decide how to use them.
• Another issue is the skills gap: many workers lack the capability to exploit AI effectively.
• Some leaders suggest hiring fresh graduates who are more accustomed to using AI, but this cannot replace internal training.
• Current investment is skewed more towards technology than people, and this priority needs to be flipped.
• According to Deloitte, 84% of workflows remain in their old state when applying AI, with only 16% designed to be AI-native.
• Redesigning work requires years, from mapping skills to identifying which tasks AI can take over.
• Most leaders predict that the workforce will not decrease significantly in the next 3–5 years, unless productivity increases are accompanied by slower growth.

📌 At the Davos 2026 Forum, the topic “why AI hasn’t had an impact yet” appeared in almost every discussion among top leaders. Many companies have forced AI usage through mandatory training or by tying it to performance reviews, but this approach has backfired. Businesses need time to train, redesign processes, and build AI-native models, rather than imposing them hastily. 84% of workflows remain in their old state when applying AI, with only 16% designed to be AI-native. Most leaders predict that the workforce will not decrease significantly in the next 3–5 years.