Workplaces are measuring AI by time saved, and missing the bigger gain
SHARE THIS ARTICLE
Many organisations are judging AI by how much faster it lets staff complete existing tasks. According to one expert, that approach captures only a fraction of the value available, because it leaves untouched the question of whether the work should exist at all.
In conversation with HR Leader, Remote’s go-to-market lead in APAC, Nick Martin (pictured), said that a lot of businesses are still measuring AI through time saved: how much faster can we write the report, respond to the email, or complete the task. “That’s useful, but it’s a fairly limited definition of productivity,” he said.
Questioning the work itself
Martin argued that many workplace processes were built around constraints that may no longer apply. “We created status meetings because people needed information. We created reports because leaders needed visibility. We created handover documents because information had to move manually between teams. AI can remove some of those constraints entirely,” he said.
He gave the example of a weekly status report; using AI to cut the time spent producing it from 90 minutes to 20 minutes looks like a productivity gain. But, if the real requirement is that a leader knows when something important changes, an AI agent can monitor the source information and post a short digest only when something material changes. “The report itself can disappear,” he said.
The question businesses should ask before automating something, Martin said, is: “Should we still be doing this at all?”
Tests for elimination
For HR leaders trying to identify which work to remove, Martin suggested a set of questions: why the process exists, who uses its output, what constraint forced it into existence, and whether that constraint is still there. Another is what the version looks like where the work doesn’t exist at all.
The strongest candidates, he said, tend to be coordination and administration rather than work requiring judgement. Examples include: recurring status meetings, information compilation, routine triage, reformatting the same information for different audiences, and reports produced because they have always been produced.
Remote, Martin said, asks employees to classify a recurring task before applying AI to it: optimise, replace, or delete.
Automation can entrench a bad process, he warned. “If nobody needs a report, automatically generating it every Friday doesn’t make it more valuable. You’ve simply made unnecessary work cheaper,” he said.
Implications for job design and workforce planning
If AI removes compilation, coordination, and routine administration, Martin said, the remaining work shifts towards “judgement, problem solving, relationships, creativity, and decision making”. This is “a shift in what people spend their time on, not simply a shrinking of headcount”, and it changes what organisations should hire for and develop.
Martin said most teams find that 30–50 per cent of what they currently do belongs on a “stop-doing-this” list, while 50–70 per cent is work AI should make faster rather than eliminate, such as judgement calls, customer relationships, building things and strategy. Workforce planning must account for both sides of that split, he said, including where people need new skills to take on the work that remains.
For HR, that means workforce planning “isn’t primarily a headcount question”. Reskilling also goes beyond tool training: employees need to learn how to redesign work, including identifying what should be automated, what should be eliminated, and where human judgement still matters. “That is a much more durable skill than knowing how to use today’s particular AI platform,” Martin said.
Shifting from visible activity to outcomes
Martin described the cultural bias towards visible activity, such as meetings, emails, reports, and busy calendars, as one of the harder parts of AI transformation. If an employee uses AI to cut five hours of administrative tasks from their week, questioning why their calendar looks less busy is the wrong response. “The point was to create that capacity,” he said.
HR leaders can help by changing what is recognised and rewarded, he said. Rather than measuring meetings attended, reports produced, or hours spent, organisations should be clearer about the outcome each role or team is responsible for delivering.
Deletion also needs to earn credit. “If the only visible win is ‘I built a faster process,’ employees will keep building processes,” Martin said. Recognition should reward someone for saying “we should stop doing this” as much as for shipping a new tool, “otherwise busyness stays the safer bet”.
The real productivity dividend, Martin said, arrives when time released by AI is redirected towards higher-value work rather than filled with a new set of tasks.
Lessons from internal adoption
Remote, he said, has adopted the principle that “work must disappear” internally, noting that the biggest lesson is that giving employees access to AI “isn’t the same as changing how work gets done”.
Remote’s AI training course is deliberately hands-on, he said, whereby participants build and ship AI projects that colleagues actually use, rather than only learning prompting or theory. The most important lesson comes after the technical training, Martin said: before building something, ask whether the underlying work should exist.
Deletion is often harder than automation, he added. Automation feels safe because the process, and often the role around it, remains recognisable. Eliminating work requires people to challenge established processes and sometimes question something another person created.
That is where HR and leaders come in. “Employees need both the skills to use AI and the confidence to challenge low-value work without believing they are making themselves less valuable,” Martin said.
The aim, according to Martin, is not to turn every employee into an AI expert but to build a workforce that is constantly asking what to keep doing, what AI can help do better, and what to stop doing altogether. “That’s where AI starts to change productivity rather than simply speeding up the way we’ve always worked,” he said.
Want to see more stories from trusted news sources?Make HR Leader a preferred news source on Google.
