Using AI for Management? AI Won’t Attend a Tribunal. You Will.

You’re running a precision engineering plant in High Wycombe or a logistics hub in Aylesbury. It’s 7:00 AM, the coffee is bitter, and your phone is already buzzing. Two people have called in sick, a delivery is three hours late, and your newly promoted floor manager—a brilliant technician who is currently drowning in the deep end of leadership—is staring at a mounting pile of staff friction he doesn’t know how to handle.

In that moment of chaos, a shiny software pitch feels like a life raft. “Automate your management,” it says. “Let AI handle the performance reviews, the scheduling, and the disciplinary triggers. Scale your business without the human headache.”

It’s a seductive promise. In my 20 years of experience, I’ve seen many “productivity miracles” come and go. So, before you hand the keys of your company culture to an algorithm, we need to talk about the “Ghost in the Machine.” Because while a computer can calculate your margins, it cannot—and should not—manage your people.

The Allure of the “Black Box”

We’re entering the era of the “Black Box” manager. This is where business owners, exhausted by the friction of “accidental managers,” delegate the hard stuff to automated systems. These systems make life-altering decisions about your staff—who gets the shift, who gets the bonus, and who gets flagged—without a shred of transparency.

Take the case of Uber drivers in the Netherlands. These weren’t “bad” workers; they were people with mortgages who logged in one morning to find their livelihood gone. No disciplinary meeting. No “quiet word” from a supervisor. Just a dead “Login” button.

They had been “Robo-Fired” by an algorithm that offered no explanation. It took years of litigation for a court to finally order Uber to explain the “logic” behind those terminations. This isn’t just a tech glitch; it’s a complete surrender of management responsibilities. Email was once the curtain struggling managers hid behind. Today, Ai wont hide. It will rule if operating unchecked.

When you remove the human from the decision, you don’t just lose “a headache”— you lose the loyalty and the soul of your staff and business.

When the Algorithm Hallucinates

Having spoken to several owners and Directors, the danger for a Buckinghamshire SME isn’t just “big tech” cruelty; it’s the fact that these systems are surprisingly brittle. They don’t think; they predict. And when they get it wrong, they get it wrong at scale.

• The 93% Error Rate: Michigan’s MiDAS system falsely accused 40,000 residents of unemployment fraud. The system was designed for efficiency, but its automated determinations were wrong 93% of the time, leading to seized tax refunds and garnished wages for innocent people.
• The Proxy Trap: A major healthcare study found that AI used “cost” as a proxy for “health need.” Because less was historically spent on certain demographics, the AI concluded they were “healthier” and denied them care. It didn’t see the person; it only saw the flawed data.
• The Reliability Gap: Even the giants aren’t immune. In early 2026, Amazon faced internal backlash after AI-assisted code changes were linked to repeated system outages that disrupted checkout and pricing. Over 1 million orders were affected.

If you let an unregulated system make “serious” management decisions, you aren’t just risking a lawsuit; you’re risking a “black swan” event that could level your entire operation – overnight.

The Rhodes & Miller Rule: Human in-the-loop approach.

I am all for innovation—my branding uses Navy (#001F3F) and Gold (#D4AF37) because they represent the high standards I believe in—and those standards require a pilot. If you are bringing AI into your workplace, you must install these three non-negotiable guardrails:

1. Mandatory Human Review: A machine can flag a problem, but only a human should solve it. “Critical paths”—hiring, firing, or denying benefits—must require a deep-dive review by a senior leader. No “rubber-stamping.”
2. The Right to Logic: Your staff deserves to know why. If an automated system flags a dip in performance on the warehouse floor, the manager must be able to explain the reasoning to the worker. If you can’t explain it, you shouldn’t use it.
3. Benchmarking the “Real World”: Don’t trust the sales pitch. Test the software against your specific sector. Does it understand the nuances of a Thames Valley manufacturing shift, or is it just applying generic logic?

The Bottom Line

Your managers are struggling because they lack confidence and leadership frameworks, not because they lack more software.

Efficiency is vital, but don’t let the pursuit of a “faster” system destroy a “fairer” one. Whether you’re running a team of five or fifty, remember: AI is a powerful engine, but it has no moral compass. It doesn’t understand “grit,” it doesn’t understand “potential,” and it certainly doesn’t understand your business as well as you do.

Don’t let your leadership become a ghost in the machine. Keep it grounded. Keep it direct. And above all, keep it human.

Thanks for reading. If you are an owner of a business across buckinghamshire, it would be interesting to hear your opinions on Ai. If you use it, what does it do for your business?. Please leave a comment below.

Warm regards,