Preparing for the AI-enabled workforce of the future

Productivity challenge

At present we believe AI will replace humans in low-stakes tasks. It is increasingly being used for customer engagement tasks, such as the pop-up web chat screens that sometimes launch when we visit websites.

But as AI becomes more widespread and demystified, and the large language models that power them cheaper to build, businesses are returning to a fundamental question - what is its value to them?  

It’s a question typically viewed through a lens of cost optimisation or productivity gains.  Measuring the productivity of modern knowledge work is notoriously difficult, but one recent study “The Effects of Generative AI on High Skilled Work: Evidence from Three Field Experiments with Software Developers”, is noteworthy.

It involved nearly 5,000 developers using GitHub Copilot in a randomised trial.

The study found that younger and short-tenure developers saw a 27 to 39 per cent productivity boost, much higher than the eight to 13 per cent boost for long-tenure developers. This aligns with previous studies showing that AI-driven productivity gains vary by skill level and education.

For businesses ready to look at their processes in a new way, the best way to assess AI’s value is the old-fashioned way, through business case assessment and return on investment.  

Opportunities for improvement need to be quantified, processes may need to be redesigned, and specific applications of AI need to be developed. Total costs, including regulatory compliance, must be measured against potential benefits.

Most importantly, people play a key role in assessing the value proposition and making AI work. Over the past year, KPMG has developed over 950 proven use cases across six sectors and 17 different functions, all of which have demonstrated business value.

People power

As part of that work several trends have emerged.

Firstly, workers still struggle with basic AI concepts and applications. Many do not grasp what AI implies for their roles, nor question why they should master a technology that might eventually take their jobs. This uncertainty underscores the need for clear communication and education about AI's personal benefits and potential.

It is also increasingly clear that the success of GenAI technologies, and the ability to realise their value, depends on a workforce’s ability to adopt and apply them effectively. Despite this, many organisations are pushing for rapid adoption before their teams are fully equipped.

KPMG’s research found that nearly two thirds of CEOs say their teams have the skills to incorporate GenAI, but few understand the impacts on the workforce. Nearly two thirds (64 per cent) say that succeeding with GenAI will depend more on people’s adoption than the technology itself.  Yet 72 per cent of organisations had not provided GenAI skills training.

As GenAI features evolve constantly, providing employees with consistent, stable, and coherent learning experiences will prove difficult.  With an ever-changing curriculum, Gen AI learning must be broad-based and continue to keep pace with change.

Employees also need abundant structured opportunities to apply and practice what they are learning. Yet AI is not well enough democratised - not every employee has access to it, or support.

This could lead to the Matthew effect - the phenomenon wherein those with pre-existing advantages accumulate more advantage over time. If access to GenAI is unevenly distributed, it could exacerbate existing disparities.

AI has already started to extend our cognitive abilities, enabling us to access, understand, and process far more information than ever before.

Highly skilled individuals find that when they explore and figure out how to use AI to support their work, it enhances and extends their capabilities without diminishing their hard-earned skills. However, for novices, an over-reliance on AI tools may limit their ability to develop essential skills such as problem-solving and subject matter expertise.

So, while Gen AI requires traditional methods of evaluating investment and return on investment, in the training and people space, we need to reconsider learning approaches.

This includes incorporating data-driven measurements such as tracking understanding and perceptions around GenAI, engagement levels, and sustained versus lapsed adoption. KPMG has been actively developing and supporting these initiatives for clients, including through our GenAI Academy.