AI improves productivity, and the current rapid advancement of generative AI is accordingly expected to increase productivity gains still further. But how will that improvement come about, and how will it manifest in the daily workflow of the people actually using and interacting with the tool?

Data from various communications platforms suggest that for office workers in particular, today's productivity losses actually come from an excessive number of online meetings and the associated burnout.

Microsoft's Work Trend Index 2023 records a 3x jump in Teams meetings within just the last year (since February 2022).

HubSpot's 2022 Hybrid Work Report found that by the end of 2021, 70% of employees had identified a high volume of calls and meetings as disruptive to their concentration and felt that the majority of meetings could have been replaced by simple emails.

But even emails take a toll on productivity. Numbers tracking Singapore workers' usage of Microsoft's Office 365 shows that people are, on average, spending 57% of their time on communication tasks – calls, emails, instant messaging, and other formats – and only 43% of their time actually getting work done.

So where does AI come in?

Thanks to the pandemic, conversations about the workplace evolved very rapidly from the hybrid model, to the use of technological tools, and now to AI, says Shyamol Bansal from Microsoft's Modern Workplace division. The reason is simple: workers are hoping that AI will take over some of these time-consuming, non-productive tasks.

“Digital debt is costing us innovation,” says Bansal, pointing to Work Trend Index figures showing that 69% of workers in Singapore do not have enough time and energy to do their job and are 3.5 times more likely to struggle with innovation and creation as a result. “Singapore employees just don't have the time to innovate or create.”

 

• 69% do not have enough time and energy to do their job.
• 82% do not have enough uninterrupted focus during the workday.
• And 70% of leaders say lack of innovation is a concern.

Unsurprisingly, this has led to a slightly paradoxical view of the technology. 67% of workers are afraid that AI will replace them – but 81% will still delegate as much work to AI as they possibly can, and not just the administrative tasks but also the analytical and creative work.

The capabilities of today's generative AI are driving those hopes, with large language models (LLMs) making the technology far more accessible to end users, and also more suited to supporting more advanced tasks – for example, generating an email response, drafting a report outline, even pulling together basic research.

“The more familiar you are with AI, the more you want its help,” says Bansal. “For most people the move to AI couldn't come fast enough.”

But how much can AI actually do?

Most leaders (30% in Singapore, 35% in the APAC region, 31% globally) say that its workplace value is in boosting productivity, followed by meeting more specific talent management needs such as wellbeing and capabilities enhancement.

In contrast, bosses who think that AI's main value is in reducing headcount are a minority (20% or less depending on region). While Microsoft's Work Trend Index does not break down these figures by industry, it's likely that these bosses come from companies which still rely heavily on manual processes or have been slow to adopt technology and automation. And, of course, the few outliers who are simply trying to slash personnel costs.

Going back to the original question, then: how should these productivity gains be achieved?

It starts with addressing the contemporary challenge of too much communication work which came about due to the pandemic, says Bansal.

“From an efficiency standpoint, you will gain a lot of time. But humans still have a role in validating the output, seeing whether it is accurate or makes sense or whether they are comfortable putting forward what the AI has created.”

Anyone can benefit from the technology, he adds – whether they are using it to drive learning, to create cross-functional capabilities, to free up time for career development, to do more value-added work, to improve wellbeing, or any other application that makes sense.

That said, he is cautious about viewing AI as some kind of panacea. While the technology can certainly help to save a certain amount of time, it depends on what the company is trying to achieve in the first place, he says.

“Each company has a different biggest disruptor. How you are going to use AI comes down to corporate culture and how is management going to address the challenge. At the end of the day you need to have realistic expectations of adoption and how much AI can do.”

What will we need to make sense of AI's capabilities?

As the technology spreads and becomes embedded in more and more common end-user applications, every employee is going to need AI aptitude, says Bansal. But the nature of that aptitude is also changing. AI skills are no longer 'hard' – LLMs have made them extremely human, starting with judgement around when to use AI versus when to use human intelligence.

 

• When to use AI and when to use human intelligence.
• Knowing how to input the right prompts.
• Knowing how to evaluate and verify AI outputs.
• Having the flexibility to integrate AI into the way we work.
• For leaders, the ability and responsibility to ensure that employees can balance business needs.

All the above skills will become commonplace soon enough, if the latest data from LinkedIn is correct: figures from LinkedIn show a 565% increase in AI talent in Singapore between 2016 and 2022, and that Singapore is one of the fastest growing centres for AI talent in the Asia Pacific.

In short, the adoption of the technology and the resulting demand for it is still on a steep upswing. There is no ideal model yet; workers are still finding ways of using it that fit their needs, managers are still figuring out how to adapt team culture and processes around it (or vice versa) and leaders are still deciding how best to put it to use for the organisation's benefit.