AI WON’T FIX A BROKEN BUSINESS. IT’LL HELP YOU BREAK IT FASTER.!

As I've said before, isn’t it astounding how rapidly we’ve all been swept into the great AI mass feeding frenzy?

Everywhere I look, organisations are announcing AI strategies, appointing AI leaders, running AI workshops and plugging AI into everything from customer service to corporate strategy.

Boardrooms that struggled to explain ChatGPT eighteen months ago are now confidently discussing “AI-enabled operating ecosystems” and “intelligent transformation journeys.”

You have to love a good journey!

Apparently, if we sprinkle enough AI fairy dust over a business, operational excellence will magically appear.

Sadly, it doesn’t work that way.

Bolting AI onto a badly run business is a little like fitting a jet engine to a broken shopping trolley.

Yes, it will move faster.

But probably not in the direction you intended—and there’s a fair chance the wheels will come off rather spectacularly.

Here’s the uncomfortable truth:

AI won’t rescue a poorly run company.

It can, however, help that company make bad decisions, frustrate customers and execute broken processes at breathtaking speed.

AUTOMATION DOESN’T MAKE A PROCESS INTELLIGENT.

IT MERELY MAKES IT FASTER!

Or, put another way: if the process is crap, automating it simply produces faster crap.

We’ve already seen some heavyweight organisations learn this lesson.

Amazon famously developed an AI-powered recruitment tool intended to identify the best candidates. There was just one tiny glitch: the system had been trained largely on historical applications from men and taught itself that male candidates were preferable. It even penalised CVs containing terms associated with women. Amazon eventually scrapped the tool. The machine hadn’t suddenly developed a personality disorder. It had simply learned from the organisation’s historical data—and then reproduced the bias it found there. AI didn’t create the underlying problem - It industrialised it. McDonald’s also tested AI-powered ordering at more than 100 US drive-throughs. Customers reported the system misunderstanding orders and adding some rather creative extras, including vast quantities of chicken nuggets and, apparently, ice cream with ketchup and butter. After several years of testing, McDonald’s ended that particular trial with IBM in 2024—although it continues exploring other AI solutions. Artificial intelligence meets the drunk, hungry customer at midnight.

What could possibly go wrong?

Then there was Air Canada. Its chatbot gave a grieving customer incorrect information about claiming a bereavement fare. When challenged, Air Canada argued that the chatbot was effectively responsible for its own information. The tribunal was having none of it and held the airline accountable for what appeared on its website.

And therein lies another important lesson: You can delegate a task to AI - You cannot delegate accountability to it.

Before pouring money into the next shiny AI thingamajig, leadership teams should ask five decidedly unsexy questions: What problem are we ACTUALLY trying to solve? Why does the problem exist?

Is the underlying process still necessary? Who owns the outcome? Should we improve it, automate it—or stop doing it altogether?

These questions may lack the glamour of an AI transformation roadmap filled with colourful arrows and clever acronyms, but they can prevent an organisation from digitally turbocharging its own dysfunction.

I’ve spent years being parachuted into businesses where people were working exceptionally hard, yet performance remained stubbornly poor.

The problem was seldom a shortage of technology.

More often, the business lacked clarity - Structures were confused - Accountability was blurred.

Decision-making had disappeared into the corporate Bermuda Triangle.

Processes had accumulated layer upon layer of historical nonsense because Brian introduced them in 2007 and nobody had dared ask why they still existed.

Customers were then forced through internal procedures designed entirely for the organisation’s convenience—not theirs.

Adding AI to that environment isn’t transformation.

It’s AUTOMATION THEATRE!

However, AI can deliver tremendous value when leadership gets the sequence right.

IKEA provides an interesting example. As its AI chatbot, Billie, took over a growing proportion of routine customer enquiries, IKEA did not simply turf its call-centre employees onto the street and declare victory for technology. It retrained thousands of them to become remote interior-design advisers—redirecting people from repetitive enquiries into more skilled, higher-value customer work. Now THAT is intelligent transformation. The technology handled more of the predictable stuff. The people applied judgement, creativity, empathy and experience. AI took part of the to-do list - It didn’t remove the human value.

This is precisely why I continue banging the drum about keeping the HUMAN IN THE LOOP.

AI is exceptionally powerful when used to eliminate repetitive work, analyse information, expose patterns and help people make better decisions.

But it still needs direction, judgement and context.

It cannot automatically understand the political history behind a dysfunctional operating structure.

It cannot rally an exhausted workforce - It cannot rebuild trust after badly managed change.

And it cannot look an anxious employee in the eye and explain honestly what the transformation means for them.

That is where the PEOPLE SIGMA comes into play—the collective trust, energy, capability, judgement and commitment that ultimately determines whether transformation lives or dies.

The organisations that benefit most from AI will not necessarily be those buying the most technology or shouting the loudest about it on LinkedIn.

They will be the ones disciplined enough to fix and simplify their businesses first.

They will remove broken processes before automating them - They will establish clear ownership and accountability - They will retrain their people rather than simply replacing them - They will distinguish knowledge from judgement.

And they will remember that technology is an enabler—not a substitute for leadership.

AI may be changing the instruments, but humans are still playing the music.

And as anybody who has heard someone attempt to play a glam-metal accordion will know, owning the instrument doesn’t necessarily mean you know what to do with it!

So, before investing in your next AI initiative, ask one very simple question:

Are we using AI to improve the business—or are we using it to avoid confronting what is actually wrong with the business?

Comments welcomed, as always!

MC

SOURCES & FURTHER READING:

Amazon’s AI recruitment tool — Reuters: https://www.reuters.com/article/world/insight-amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK0AG/

McDonald’s AI drive-through trial — Associated Press: https://apnews.com/article/bebc898363f2d550e1a0cd3c682fa234

Air Canada chatbot ruling — American Bar Association: https://www.americanbar.org/groups/business_law/resources/business-law-today/2024-february/bc-tribunal-confirms-companies-remain-liable-information-provided-ai-chatbot/