Technology   //   January 11, 2024

How AI ‘skill leveling’ will arm workforces in this year’s ‘quest for productivity’

Never has the expression “tool up” been more pressing.

Some may even say urgent if the rhetoric around generative artificial intelligence’s potential is to be fully realized in all desk-based jobs. And with productivity such a keen watchword for 2024 among business leaders, figuring out how to skill up the right way in AI, is top of mind for many. 

For senior business executives, important strategic decisions are still to be made this year around what parts of their businesses will benefit from AI, what type of AI to invest in, how to do so ethically and with which vendors, and what goals should be set to establish a return-on-investment plan. For employees, it’s all about the training. 

Much as there is a lot of interest in how much AI may lessen employee workloads, there is still latent anxiety when it comes to its application at work, as recently outlined in a report from accounting giant EY. Over two-thirds (67%) of 1,000 desk-based or office workers in the U.S., said they’re concerned they will lose out on promotions if they don’t know how to use AI technology and 66% said they’re worried about falling behind on work if they don’t use AI. 

Business leaders share that worry. More than half of business leaders surveyed by YouGov on behalf of Microsoft in the U.K. said that they’re concerned their workforce lacks the skills to make the most of the AI opportunity. While only 25% said they have plans in place to recruit the right talent to successfully implement the tech across their organizations, per the same research. And just 26% have completed training to improve their understanding of how to use AI in their jobs. 

Helping convert the AI aspirational rhetoric that businesses have in countries like the U.K., and the U.S., with tangible business plans, is a goal for tech suppliers like Microsoft, which has pledged to train 1 million people in the U.K. on AI. “The starting point for all of this is the culture that you have within your organization,”: said Simon Lambert, chief learning officer at Microsoft. “To foster an AI-ready culture you need to have a culture of learning because things are moving so fast.” 

That means giving people the space, resources, guidance, security, and support to think about how they improve their own capabilities around this new technology, stressed Lambert. “They also need to create time for learning. Organizations should be helping craft time for people to take what they need to go and build the skills they need. And the third thing is [to make] room for experimentation. Because when you’re going on this learning journey, you need to celebrate and acknowledge success, but you also need to recognize when things could go better.”

Rise of “skill leveling”

We’ve all heard of reskilling and upskilling, but what about side skilling and skill leveling? These latter two are now being referred to by tech experts to describe how desk-based workers will need to hone their AI skills in the coming year. They go hand in hand with the concept of the augmented workforce, aka the kind of workforce that is made more efficient, therefore productive, by the smart application of AI tools.

“The old philosophy that we want to try to get the best out of our humans and the best out of our technology and have them work together has suddenly materialized as a practical opportunity for organizations,” said Dr. Chris Brauer, director of innovation at the Institute of Management Studies, Goldsmiths, University of London, where he conducts large-scale research projects on the intersections of emerging tech and the future of work and human behavior. “There are decisions that can be taken that can create that alignment between your human workforce and your AI resources and capabilities that then increases the productivity of the broader firm as opposed to just the individual workers,” he added. 

Brauer says there are several kinds of strategies that organizations can use to ensure their workforces develop the necessary skills to make the best use of AI. First up: skill leveling. “This is where low-to-medium-skilled workers benefit disproportionately from using generative AI augmentation than highly skilled workers,” he said. 

“We can imagine lots of things they would do, but in practice, it hasn't materialized and very few organizations have strategies on what they would do with that extra time that's freed up from the expert and highly skilled workers so that’s what we need to see evolve in 2024.”
Dr. Chris Brauer, director of innovation at the Institute of Management Studies, Goldsmiths, University of London.

In all early-stage experiments run at Goldsmiths, there is nearly a 40% increase in the average quality of output when the AI is given a specific task, he stressed. “You give the task to the low-to-medium-skilled desk-based worker without the AI, then you complement them with the AI and you look at the output that they’re able to produce – things like, write a press release for example – a task that if left to their own devices, they can produce to a certain level of quality as they’ve got low to medium skills and experience in the area,” he said. “The highly skilled worker who’s not augmented can produce a much higher quality output typically but when you augment the low-to-medium-skilled worker with the AI, the quality of the output actually exceeds the augmented, highly skilled worker.”

That means that by using generative AI tools, this level of worker can jump up to the level of a more senior, experienced worker much faster – raising the overall bar for the company in general. 

The results for those more senior workers are quite different. “When we augment the highly skilled worker, the bump is much less, it’s incremental – more in the range of a 10% improvement on the quality of the output because they’re already capable of producing a very high-standard quality of press release in the first instance,” said Br

auer. That smaller bump, however, does still take them above the augmented [by AI] low-to-medium-skilled worker. 

So a business can identify what all the business processes and routine tasks are that require certain outputs, and then train this level of worker to carry them out using generative AI, and in doing so rapidly upskill them. They’ll therefore need to understand prompt engineering, which will enable them to ask the right sort of questions to the AI system about the press release, to improve it, and to iterate it into a higher quality version. 

Masterclasses in prompt engineering are becoming increasingly accessible, and only take a few weeks. “My students are highly proficient in the art and science of prompt engineering at the end of a two-week masterclass at the University at Goldsmiths,” he added. “And that makes them capable of being that low- to-medium-skilled entry-level worker who then, complemented by the [AI] system, is suddenly producing an output that’s 40% better than it would have been if they didn’t know how to use the system effectively.”

Once they’ve had that training, they then have to work with the AI as their so-called copilot (what some are calling side-skilling) to maintain this level of efficiency and outcomes. 

So far so good. Upskilling the more experienced layer of desk-based workers, however, is a grayer area. These so-called “pro-workers” as Brauer refers to them, can use AI tools to perform the less-skilled, more routine, and time-consuming tasks and in doing so free up to 30% of their time. Figuring out what they do in that 30% of freed-up time will be a priority this year, stressed Brauer. 

“That’s where we don’t yet have a lot of answers because there haven’t been many good demonstrations of organizations that have the foresight to anticipate: if we free up our expert workers to do these higher-order tasks, what are they [those tasks]? We can take a chief operating officer and free up 30% of their time by smart augmentation with an AI system. Great. But they were fully booked with their existing activities. These are new, fresh activities, that they’re freed up for – not necessarily more of what they already do.”

It might be that they can now spend that time developing new product ideas, innovating within the organization, and applying creative and critical thinking to complex challenges that the organization faces – all tasks that wouldn’t have been possible before when they were bogged down in operations. “We can imagine lots of things they would do, but in practice, it hasn’t materialized and very few organizations have strategies on what they would do with that extra time that’s freed up from the expert and highly skilled workers so that’s what we need to see evolve in 2024,” added Brauer.

Getting buy-in from employees 

The loftiest technological aspirations, or the deepest investment pockets a business can have, won’t count for much if employees aren’t brought along for the ride. 

There is evidence to suggest that when technology is implemented by businesses that haven’t communicated about it transparently with the workforces that are expected to use it to reap efficiency benefits, the results can be disastrous. 

It wasn’t so long ago that digital transformation (DX) was the favored buzzword within organizations, rather than AI. But one in two of these DX projects failed globally in the past year and that has had a direct impact on staff attrition, found a report by tech firm Endava published last month. 

The Leveraging the Human Advantage for Business Transformation also highlighted that staff feel frustrated (56%) after projects like this fail, and are likely to leave (50%).

A core reason why such projects have failed is often tied to a breakdown in communication, stressed Endava CTO Matt Cloke. In some cases that’s when technological teams creating new systems aren’t getting feedback from teams who are executing the work, what tech would make their jobs easier and more efficient. So that means that even if an end product has all the technological bells and whistles, it may wind up on the scrap heap if no one internally actually uses it. 

“There is this kind of quest for productivity. Businesses are focused quite deeply on it. So it's not about cutting heads, but about how can we do more with what we have, as opposed to necessarily hiring more people.”
Matt Cloke, CTO, Endava.

Or it can simply be that after huge investment in the tech, senior leadership fails to communicate why it’s being rolled out, so staff are non the wiser and default to former processes. 

This “employee engagement problem” will remain as businesses race to integrate AI if they make the same mistakes and don’t communicate their strategy effectively to employees, he stressed. “Where we’ve delivered [digital transformation projects] and we’ve seen that kind of bad behavior, there seems to be a – let’s call it stress level – that exists in the senior parts of the organization whether it’s a crunch on profits or revenues, or there’s a driver that is causing stress at this type of level. But the right articulation doesn’t make it down inside the organization,” said Cloke. 

This year, businesses are regrouping after the subdued financial year most had in 2023, and leaders are occupied with attaining high productivity, said Cloke. “There is this kind of quest for productivity. Businesses are focused quite deeply on it. So it’s not about cutting heads, but about how can we do more with what we have, as opposed to necessarily hiring more people,” he added.