Study: Why upskilling in generative AI increases productivity by over 150%



For leaders, the case for integrating generative AI into working practices is compelling – and it seems it’s equally vital to invest in upskilling solutions to help teams reach the cutting edge of skills knowledge as quickly and effectively as possible. But the impact of Gen AI on how teams execute their roles depends on a number of factors:

• Which tasks are being automated or enhanced using generative AI like large language models (‘LLMs’) as co-pilots? 
• Who in the workplace is most affected by these changes and what are the short and long term implications for their careers? 
• Who makes the decisions about how and where Gen AI is being used in the workplace, and what best practices should be established to ensure the responsible use of generative AI usage?

According to McKinsey, current generative AI and other technologies have the potential to automate work activities that absorb 60 to 70 percent of employees’ time today. This is a sharp uptick from McKinsey’s own predictions in 2017 that generative AI technology has the potential to automate half of the time employees spend working.

McKinsey also predicts that banking, high-tech, and life sciences are just three of the industries that could see the biggest impact in their revenues from generative AI use in the business. For example, across the banking industry Gen AI could deliver value equal to an additional $200 billion to $340 billion annually at full implementation levels. In retail and and consumer packaged goods, the potential impact is also significant – an estimated $400 billion to $660 billion a year.

Which tasks are more productive when using GenAI?

We surveyed data practitioners working in a variety of roles such as Machine Learning Engineer, Data Scientist, AI Scientist, Software Engineer, Chief Information Officer, and CEO. The full results are detailed in Workera's latest ebook, 'The Leader's Guide to Generative AI Skills in the Workplace' which can be downloaded here for free.

We asked them to estimate how their productivity has improved or will improve when using generative AI as an input source or co-pilot. Respondents were asked to assume a task currently took them 10 hours to complete without any input from generative AI, and then to estimate how long a task does take or would take to complete from zero hours (where generative AI enables the task to be fully automated) to more than ten hours (where generative AI actually increases the amount of time a task would take to complete).

The tasks we asked respondents to estimate an improvement in when using generative AI as an input source or co-pilot:

• Language translation
• Documentation generation
• Text analysis
• Drafting emails and correspondence
• Algorithm exploration 
• Code generation and autocompletion
• Simple data analysis
• Data preprocessing
• Idea generation
• Code debugging assistance

Our survey respondents reported that, on average, tasks that would take them 10 hours manually now take them (or could take them) between five and six hours less.

We found that text-heavy tasks like language translation, text analysis, documentation generation, and email drafting were the four tasks with the biggest increase in productivity. Assuming these tasks would have originally taken 10 hours to complete without generative AI, the respondents who are already using generative AI such as LLMs to help augment or co-pilot their work reported experiencing an average 168% increase in productivity.

“Respondents who are already using generative AI such as LLMs to help enhance or co-pilot their work reported experiencing an average 168% increase in their productivity.”

For the more complex tasks in our survey, this was still an average of over half as fast as manual practices. This adds up to a 150-160% increase in productivity for all the example tasks we included in our survey, indicating that teams could achieve results twice as efficiently as they currently do when using generative AI to enhance their workflows. The full data can be found in the free download here.

What this means for leaders

According to McKinsey’s updated adoption scenarios, half of the work activities currently happening in organizations “could be automated between 2030 and 2060, with a midpoint in 2045” – a decade earlier than McKinsey’s previous estimates. This is likely accelerating thanks to a potent combination of “technology development, economic feasibility, and diffusion timelines”.

McKinsey surmises that “generative AI can substantially increase labor productivity across the economy, but that will require investments to support workers as they shift work activities or change jobs.” For leaders, this means supporting their teams to upskill rapidly, sustainably, and comprehensively.

Workera's free guide explains how and why leaders can support their teams in upskilling in generative AI – ensuring business transformation keeps pace with the latest in GenAI innovation. Download the guide for free here.