Employee wellbeing has become a central focus for businesses during the COVID-19 crisis, with the pandemic putting the spotlight on the need to promote healthy working habits and ensure staff are provided with the support they need to maintain good mental health while working far from the office and colleagues.

While many companies may have attempted nothing more sophisticated than regular team catch-ups on Zoom to keep an eye on things, some have turned to more novel approaches to gain insight into how staff are faring during difficult times.

SEE: Managing AI and ML in the enterprise 2020: Tech leaders increase project development and implementation (TechRepublic Premium)

For the past three weeks, professional services firm PwC has been running a pilot scheme that combines machine learning with wearable devices to understand how lifestyle habits and external factors are impacting staff. The project involves volunteers using fitness trackers that collect biometric data around the clock which, when combined with a series of cognitive and biometric tests and fed through an AI algorithm, aims to help staff better manage stress.

The key to this initiative is observing how factors such as sleep, exercise and work diary load play a part in how PwC staff perform at work. Eventually, it is hoped workers will gain a better understanding of how they can structure their approach to work and home life in a way that benefits their mental health and wellbeing.

It’s a theme that resounds with many of us currently, which perhaps explains why the initiative has proved so popular with PwC employees: the company saw 2,000 applicants sign up within four hours of posting the project on its internal intranet for a trial with only 1,000 spaces.

“I think if we went higher profile, we could have had a multiple of that number, which starts to get up to a pretty high percentage of the organization in the UK,” says Euan Cameron, PwC’s UK artificial intelligence lead.

“What is says to me – and it’s really the first learning from the exercise – is that there’s really no shortage of interest… we would never do this if an organization tried to compel people to do this. I think you’d be in different territory, frankly.”

The pilot scheme at PwC came about following discussions between Cameron and associates at IHP Analytics, a boutique analytics firm that specializes in human performance in elite sports. The firm, which has worked alongside professionals in Formula 1 racing and Olympic cycling, is aiding the development of the underlying platform, which it eventually hopes to offer to external clients.

“One of the areas, even before COVID, that we knew was developing fast was a deeper understanding of human performance and human wellness,” Cameron says. “We want to marry these two together to do something positive for our people.”

Vicki Broadhurst, a senior manager at PwC, volunteered for the trial in order to help her understand how her physical activity linked to her cognitive performance and how she felt.

She tells TechRepublic that her participation in the trial stemmed from her own interest in the role of artificial intelligence in psychometric testing, as well as wanting to remain active during lockdown.

“I wanted to take part in something that would challenge me to be more active whilst I was at home all the time, as well as give me targets to work towards,” she says.

“As a morning person, it’s made me realize that if I don’t get my steps in early in the day I’m less likely to do them later. This means that I now structure my day around regular early breaks to do exercise, which makes me much more productive overall.”

Crucial to the fitness tracker – a Garmin Vivosmart 4, in this instance – is that it can capture heart rate variability, or HRV. Cameron refers to this as “the Rosetta Stone” in biometrics, insofar as it is able to record changes in the wearer’s pulse rate associated to stress, energy and recovery – “which is really important for what we’re looking at here,” he says.

In addition to wearing the fitness tracker 24/7 (Cameron stresses that participants are allowed to take them off to shower), employees are asked to perform a series of cognitive tests, either delivered via online games or short questionnaires.

Performed daily, these 10-minute assessments have been designed to evaluate individuals’ short-term memory, concentration, and task-switching ability, to help them understand how on the ball they are, from a cognitive point of view, each day.

This is combined with a one-off set of more in-depth baseline tests, which aims to provide insight into “some deeper personality traits”, Cameron explains, including risk appetite and working preferences.

SEE: Guide to Becoming a Digital Transformation Champion (TechRepublic Premium)

Finally there’s the contextual data. This looks at factors such as how heavy an individual’s workload is on any given day, how many hours they’re working, and even their home working environment: “How fast is your broadband, have you got a dedicated space where you can work, are your kids at home, things like that,” says Cameron.

“It’s quite a rich dataset, which will help us understand what’s driving performance at an aggregated level. But primarily, [it’s] to help individuals understand a little bit more about their mental and physical response to the world of work, and the different stimuli that exist.”

This in turn provides an opportunity to learn “something about how we cope as individuals, and as an organization,” says Cameron.

Cameron isn’t in a position to share any data insights from the trial yet – it’s early days, after all – and the data needs to be “stress-tested, poked and prodded,” he says. However, he does suggest that early insights point to an “inflection in energy and stress levels” when staff have more than seven appointments in a day.

For Broadhurst, the most interesting insights have been those into her sleeping patterns, as well as the more subtle nudges the smartwatch delivers to encourage her to move around or stay hydrated.

“Stress levels are not something that I have been particularly focused on as I am generally quite a relaxed and calm person,” she explains.

“I’m more interested in things like my sleep patterns, and the device can help with that as it breaks them down to light sleep, REM sleep and deep sleep. I discovered that I’m getting more sleep in general than before, but not as much deep sleep.

“I have also done some baseline cognitive tests, and we should get the results soon, which I’m looking forward to.”

Equally interesting for Cameron is the variation in individuals who underestimate or overestimate their stress levels, accordingly. “Some people are exceptionally good at being in-tune with their bodily response, whereas other people who find it quite difficult,” he says.

“It’s prompted me to look at alternative ways to manage my energy levels, manage my stress levels and think about relaxation, other than sleep and exercise outlets. I think a lot of the benefit of this from an organization… comes from a thousand small changes that come from the grass roots, on the basis of the information and insight you put in peoples’ hands.”