Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Feeling like a cog in a wheel — loaded with tasks yet powerless to decide how and when to do them — has long been known to be stressful and unhealthy.
The landmark studies on British civil servants, led by Sir Michael Marmot over several decades, even found a link between a lack of autonomy and heart disease.
Yet there is no shortage of automated and prescriptive protocols in the modern workplace — with the spread of AI likely to also reduce some workers’ autonomy. So is it even possible for employees today to enjoy having autonomy and the satisfaction of solving problems creatively?
The meaning of autonomy can vary between people and over time. “If you think back pre-Covid in our business there was no autonomy over where you worked, you worked in the office,” says Ian Nicholas, global managing director for the recruiter Reed. Today, Reed’s consultants can work flexibly between home and office, and argue as much about how, as well as when and where, work is done.
One example is a piece of software that Reed asks its recruitment consultants to use. Powered by machine learning, it can rapidly turn out job adverts with just the right bullet points and keywords to secure the top spot on job boards.
But not all staff love it. “We do have a lot of long-serving consultants and if they don’t like the way the advert has been phrased, they want to change it. At the moment, we just let them,” explains Nicholas. “There’s some in my business that would say that’s silly, the tech does it in the best way, we should let the tech do it. But I think that makes very experienced people get very frustrated, and I worry they would just fall out of love with the job.”
There is also the added risk that, if such software becomes commonplace, adverts “are all going to start looking the same”, Nicholas reflects.
AI automation “works best when it removes friction and difficulty”, says Neil Sawyer, HP Northern Europe managing director. By lightening the load — handling rote inquiries, analysing data — “you free up valuable human resource to deal with those more complex issues”, he adds. Yet that is not the whole story.
For a start, technologies are rarely flawless or even designed to be. Most companies “accept that an automated process is going to work about 80 per cent of the time” because they cannot fully customise the solution they purchased, observes Caroline Hughes, chief executive at Dublin-based leadership adviser Conscious Leadership Development. Left unchecked, automated systems can cause errors and make correcting blunders harder. They remove “the very discretion that enables people to do their best thinking”, Hughes adds.
AI and algorithmic management has revolutionised many “gig economy” jobs — taxi and delivery drivers are now often directed by software rather than a human manager. This is now advancing on white-collar occupations. Typically bundled into enterprise software, the latest tools break jobs down into simple tasks and metrics, enabling individual performance to be quantified and compared, says Steven Rolf, a researcher at the University of Sussex’s ESRC Digital Futures at Work Centre.
Yet where skilful managers take a “well-rounded view”, algorithms stick to what is measurable — calls handled, orders processed — disregarding all the other things, such as mentoring colleagues or taking ownership of a customer’s intractable problem. Marketed as boosting productivity, the technology risks degrading it as managers feel obliged to accept “computationally derived” assessments of staff and workers put hitting targets before service quality, Rolf adds. “There’s an assumption of expertise, because [the tech] was built by world-class Silicon Valley engineers.”
As technology advances, workers need to “see a future for themselves”, says Debra Maxwell, chief executive of customer experience business ArvatoConnect. To instil belief that they themselves can build that future, the company mostly selects managers from its ranks. It also tries to make employees feel that they can make their jobs better. To this end, Hughes recommends feedback channels, or “fix-it-forums”, in which employees repair faulty processes. Individually small, the tweaks — streamlining approvals protocols, de-duplicating workflow steps — add up.
That said, fixing has drawbacks. Oliver Shaw, chief executive of organisational design and workforce planning business Orgvue, says: “If you create a tolerance for high error rates, you’ve made it OK for stuff to go wrong,” stressing everyone.
Given the need to expose vendors’ hype yet build the AI competence of staff, employers might do well to give workers more say in designing systems, suggests Rolf. They, after all, understand “the bundle of tasks” that makes up each job, what it makes sense to automate and where a person should decide.
Using technology to expand not prescribe choices is something Reed works hard at. “We often teach our consultants you might shortlist three people — but then throw in a rogue candidate that doesn’t necessarily meet the spec, because you just have a gut feeling that person might be right,” says Nicholas.
Read the full article here