Behind the rosters, compliance checklists and client reports that keep a modern cleaning or facility management operation running, a contradiction has emerged among the very people relied upon to keep things spotless and above board. A new Employment Hero report has found that, although Australian workers are among the fastest adopters of artificial intelligence in the world, close to half of them feel guilty about using it, and more than a third admit to hiding that use from their employer altogether.
For an industry built on trust, supervision and demonstrable standards, that finding hits hard. Cleaning and facilities management businesses depend on frontline staff, supervisors and back office teams to accurately log hours, draft reports and manage client communications, and AI tools assist with all three. Yet the research suggests many workers regard that assistance as something closer to cheating than to competence, with 42 percent of respondents describing AI use at work in exactly those terms.
“I reckon there’s more AI use happening than managers realise,” The Robot Factory founder Tom Culver says, “particularly for everyday admin like drafting emails, preparing quotes and tidying up reports. I can’t speak for every frontline team, but those are practical uses that can save people time and help them get their point across.”
It’s not the productivity case that’s in doubt. Three in four Australian workers reported an improvement in productivity as a result of using AI, and a near identical proportion said it improved the quality of their output. Among cleaning and FM operators already stretched by labour shortages, rising compliance obligations and thin margins, those gains make a big difference. The redirecting of staff saved time for more strategic or higher value tasks, as 29 percent of respondents claimed, and is precisely the outcome operators want from any new technology.
Industry figures confirm the pattern is already visible on the ground. Big Clean operations manager Julie Wang says informal AI use is common among customers and staff alike, even where it sits outside official business practice. “What we hear from our customers is that they are using it for various levels of admin, on everything from drafting emails to updating cleaning schedules.” Wang adds that, in store, customers increasingly turn to AI on their phones to navigate product information and label translations, a task they would previously have brought to Big Clean staff directly.
HR On Call director Melissa Behrend cites a simpler explanation for the secrecy uncovered by the research. She says many workers are reluctant to disclose their AI use because their employer has no formal policy on the subject and has communicated nothing about acceptable use, leaving staff unsure whether admitting its use carries any risk.
Why the guilt persists
The disconnection, according to Employment Hero chief executive Ben Thompson, has little to do with the technology itself and everything to do with the culture surrounding it. Employers who fail to set clear expectations leave staff to work out the rules on their own, and that uncertainty breeds concealment. Thompson argues that businesses building AI literacy into onboarding and normalising open conversation about its use will pull ahead of competitors, not because their people are more willing to try AI, but because they already are using it and simply need permission to continue.
Wang identifies a further, more concrete danger sitting alongside that cultural gap. She warns of staff loading documents into personal AI accounts with confidential client information, such as alarm codes, after-hours access, floor plans or client staff routines, saying that this gives businesses no control over where that data is stored or who ultimately sees it.
Culver adds to that worry. “The concern for me is what information people put into these tools – client details, site security information or confidential contract pricing. There’s also the risk of an AI-generated answer sounding right when it isn’t, especially with safety procedures or compliance, so someone who understands the job still needs to check it,” he says.
A second and less obvious hazard that Wang raises is when AI drafts technical documents. “Businesses use AI to draft critical workplace health and safety (WHS) documents like SWMS or MSDS, which may save time and read reasonably, but may actually be factually wrong and become a real safety issue for users.” Errors of that kind, she notes, could mean the wrong chemical dilution damaging floors or unsafe handling instructions putting cleaners at direct risk, with the business ultimately carrying responsibility for the outcome.
Behrend raises a related concern about staff who use AI tools without fully understanding what happens to the information they share. She says the danger lies in users not grasping the parameters and limitations of the tools they rely on, which creates a genuine possibility that sensitive information entered into an AI system could later be accessed by others.
Building trust into adoption
Both commentators point to policy and communication as the practical fix. Wang recommends that operators should assume AI use is already occurring and work with teams to understand where and why it happens, before moving staff onto business accounts with appropriate security settings and clear rules on what information can and cannot be entered into a platform. She stresses that anything client-facing or safety-related created with AI assistance should always receive a human check as a final step, and that customers unsure about product use should default to asking staff directly rather than relying on an AI response that may contain errors.
Behrend’s advice centres on the same starting point. She says businesses need a clearly communicated AI policy that spells out what is acceptable and what is not, rather than assuming staff already understand the privacy and confidentiality implications of the tools they use. Where a company has a preferred enterprise AI option, she adds, all employees should be directed towards it.
“Give people clear guidance on which tools they can use, what information they can share and what needs checking,” Culver adds. “If someone finds a way to save an hour on paperwork, encourage them to share it rather than feel they have to hide it. We want people using their judgement and spending more time on the work that needs them.”
The lesson for the sector is straightforward: productivity gains delivered from AI are only as valuable as the trust and governance surrounding them, and cleaning and FM businesses that get ahead of that conversation stand to gain both the efficiency and the workforce confidence that guilt-driven concealment currently puts at risk.
