AI being mandatory in businesses without rules is “like handing out power tools with no safety briefing”, experts warn.
Many businesses are requiring staff to use AI at work to meet efficiency goals, ensure consistency, and prove return on investment (ROI).
But small medium enterprises (SMEs) are “just speeding up mistakes”, experts have warned.
McKinsey data found 22% of employees say they’ve had none or minimal support to build AI capability.
Kate Underwood, Founder at Southampton-based Kate Underwood HR and Training, said using AI in businesses can lead to mistakes.
She added: “Calling AI ‘mandatory’ without rules is like handing out power tools with no safety briefing. Yes, you can require people to use AI at work. But for UK SMEs, ‘just use it’ is how you end up with client info pasted into the wrong tool, AI-written waffle sent to customers, and a complaint landing on your desk. And guess what?
“The business owns the mess, not the employee you pressured to hit ‘generate’. Forced AI also breeds fear and fakery. People either over-trust it, or they’ll use it in secret on personal devices because it feels easier than asking for help. That’s not efficiency. That’s the risk with a spreadsheet.
“If you want ROI, do the boring bit first: decide what AI is for, what’s banned, what data is a hard no, and make ‘human check before it leaves the building’ non-negotiable. No policy and no training equals no mandate. Simple as that.”
The business owns the mess
Colette Mason, Author & AI Consultant at London-based Clever Clogs AI, said we need to remember AI gets it wrong “30% of the time”.
She continued: “Companies are drowning in toxic AI hype from vendors and social media gurus promising miracles, then terrifying staff with AI workflows before anyone understands what the tools do or where the guardrails are. Workers don’t want to embed AI tools and end up with a P45 as thanks. No one wants to say they’re even using AI.
“Companies need to know what AI can legitimately do well right now, with proper guardrails. Notetaking, summarisation, brainstorming, ideation using GDPR-compliant tools like Microsoft Copilot is a genuine productivity boost, not science fiction. Automation based on logic is an absolute asset, though marketing blurs that line.
“AI still gets it wrong 30% of the time (even with AI checking AI outputs). “If this happens, do that” logic is safer, eg if a client books a meeting, automate the reminders. But the moment you let AI make customer-affecting decisions without a human in the loop, you’re violating both regulatory compliance and basic customer service principles.”
AI still gets it wrong 30% of the time
Rohit Parmar-Mistry, Founder at Burton-on-Trent-based Pattrn Data, said the trust of clients could be destroyed.
He added: “Mandating AI usage to hit efficiency targets is the fastest way to scale incompetence I’ve seen in a decade of automation work. Leaders think they’re driving innovation; actually, they’re just pressing the accelerator on mediocrity. The McKinsey data showing that 22% of staff lack support isn’t a skills gap, it’s management negligence.
“If you force employees to use ‘black box’ tools without defining what good looks like, you don’t get faster work. You get plausible-sounding rubbish that destroys client trust in seconds. I’ve spent 10 years fixing corporate automation messes, and the pattern is always the same: leaders hand out power tools without a safety briefing and then blame the workforce when injuries happen.
“The real danger isn’t that your staff won’t use AI; it’s that they’ll use it to bypass thinking. Don’t measure AI adoption by login frequency. Measure it by whether it actually makes the human’s job better. If it requires a mandate, your implementation has already failed.”
You get plausible-sounding rubbish
Mark Hosker, Mortgage Adviser at Bradford-based Cyborg Finance, said it’s more about how we use AI than it being inherently negative.
He continued: “Disregarding isolated lapses of judgement, errors and complaints arise not from the AI, but from its inappropriate application.
“With employees, this highlights the need for enhanced staff training in the proper use of AI within the organisation. Adopting a mindset that treats AI as a junior trainee who requires active supervision and oversight provides a strong, practical foundation.”
Photo by Kenny Eliason on Unsplash.
