Is it "Awright, pal?". Or is it "Awright, pal?"

I'm proud to come from Glasgow, even though its more than 30 years since I lived there.

In my hometown, the expression "Awright, pal?" can mean one of two things.

It can be a friendly greeting to a stranger who turns up at the bus stop where you're waiting for a bus.

Or those can be the last words you hear before a full-blown riot breaks out.

The difference is context.

It's a bit like that much-shared stat about how communication is only 7% the words we say. 38% is our tonality - how we say what we say. And 55% is our body language and facial expressions.

Believe me, if you're talking with a Glaswegian, you'll know whether they're being friendly or seeking to start a riot from the 93% of their communication that goes beyond the specific words they're speaking.

It's the principle that counts

Now, there are some people who argue with the precise numbers above, and that's fine. I'm usually sceptical of numbers developed in the social sciences too. They're often a partial and unrepresentative perspective on an issue.

So it's easy to spark an argument about whether the "words-alone element" is 7%, 12% or some other number.

But that's to miss the point.

The only point that matters is the principle that the precise words spoken are a small part of the total communications package that reaches the person you're talking to.

Whether that additional impact is a precise 93% of the total communication or some other number, it's large proportion of it, for sure.

And if you doubt me, consider this: for most of us growing up, the way our parents said our names would be an indication of how angry they were with us.

That's just our name. A single word.

Imagine what it's like when a communication is several sentences long. Or a paragraph. Or a book.

Dame Edna Everage

I've always been a huge Dame Edna fan (less so, Sir Les, but that's for another day).

Many years ago, I bought Dame Edna's autobiography and was reading it on the London underground one day.

I was slightly disappointed, to be honest, as the prose felt a little flat.

But, a few pages in, for some reason I started reading the book inside my head in the voice of Dame Edna's character and it immediately turned into the funniest thing I've ever read.

(I can only apologise to any fellow commuters from that day who might have been scarred by the experience of a large Scotsman breaking out in regular loud guffaws as he ploughed through a book, tears of laughter streaming down his face. That sort of behaviour is generally frowned upon by London commuters.)

The words were the same.

Reading it in my voice inside my head made even double-entry bookkeeping seem exciting.

But reading it in Dame Edna's voice, I wasn't sure I'd be able to control my bladder until I reached my stop.

Same words. Exactly the same words. Entirely different impact.

Your bottom line

In case you think that's all very interesting as far as it goes, but has nothing to do with your business, try this.

Imagine your profits went up by 10% last year.

That's a cause for celebration, right?

Well, not if everyone else in your sector doubled their profits.

What about if your profits fell by 10%?

Terrible news, surely?

Well, not if the rest of your sector plunged into billion-dollar losses.

You see, shorn of context, no number, no set of words, no "fact" means anything.

Whether it's 7% of the message, 12%, or some other number, context - or tonality, expression, and body language, if you prefer - accounts for the vast majority of the impact of any piece of communication.

Unless you're a tech bro

Unless you're a tech bro, that is (or tech gal, for that matter).

Because this is a group of people who think only in terms of writing a precise line of code to achieve a micro-outcome on the path to, say, making a printer produce a piece of paper with the correct image on it, with rare exceptions, they think only the precise wording matters.

To be fair, if you're trying to make a printer work, that's probably true.

But when you're trying to interact with humans...and indeed humanity itself...things get a bit more complicated.

If "just the words" was all that mattered, then we'd all be delighted with the website chatbots we need to negotiate before we can get a customer service issue fixed.

We wouldn't mind the cack-handed website design which means we have to enter the same information more than once, or in an illogical order, to navigate through some process or other another.

We wouldn't even mind the fact that you sometimes need a PhD in computer science just to reset the password to access a website you've visited to hundreds of times before.

But none of that makes anybody happy.

The words are all that matters...allegedly

Mainly because I try to watch my language on this platform, this isn't the place to tell you how I felt about the website which asked me to reset a password.

Which I did.

Only to be told, after I'd entered it twice and pressed "confirm" that I hadn't used a capital letter, which was mandatory.

I figured that might have been more usefully explained before I'd entered my new password than after, but hey-ho, I thought.

Mildly irritated, I made one of the letters a capital, then repeated the process. Only to be told that I hadn't used a number in my password, which was also required.

Now, quietly fuming, I entered a number on the end and repeated the rigmarole once more.

Only to be told that my password also required a special character like # or * as well, or it wouldn't be valid.

My reaction on the third time this error message came back to me?

Let's just say it was one of the possible meanings of the Glaswegian expression "awright, pal?".

But think about what happened here.

At each stage, the error message I received was factually correct, according to the definition applied to password construction by the business I was going through this process with. There was nothing wrong with their communication of the facts.

They correctly identified the factual errors I'd made in constructing a password according to their specifications.

But how they communicated those facts was a masterclass in how to make sure all your customers hate you.

The 7% factual information was accurate.

The lack of consideration for the 93% of the message - the context, if you will - made me furious, which generally isn't how you want your customers to feel about you.

What sane person, thinking about the human user, wouldn't have put the full password criteria just above the box where you register your password, rather than drip-feed out each different requirement one at a time.

And that's my problem with AI

There's no doubt AI is a very clever idea, developed by some very clever people.

There's equally no doubt in my mind that it was developed by people with a deep and abiding interest in the 7% of the communication process which is the factual stuff, and a complete lack of understanding - or even awareness - of most of the 93% of the messy, complex, human component.

If you think I'm being a little harsh in my judgement, the other weekend I saw an article about a competition for building an "AI girlfriend" which was all kinds of creepy and disturbing.

No doubt the factual challenge writing the code was considerable. (7%)

But it clearly crossed nobody's mind that the whole process was degrading and dehumanising to 50% or so of the planet. (93%)

I'm sure AI will turn out to have some uses.

Could it handle factual challenges, like maths computations, to a moderately competent level? I'm sure it could.

Might it be helpful to sift through a huge pile of factual documents and summarise them? Possibly.

Is it possible that AI could tell me how many times the word "encounter" was used in business pages of the New York Times between 1987 and 20003? I'm pretty sure it could.

But even if AI reaches perfection, we can only expect it to address a maximum of 7% of what we need to know to make good decisions.

Its lack of humanity - a missing link that will never be filled no matter what the tech bros and tech gals tell us - means that it will always be blindsided by the 93% of the communication it doesn't understand and never will.

So if you're a big company obsessed with how AI is going to transform your business, remember it will only transform a maximum of 7% of it, because it only does facts and can't understand the context that represents 93% of everything you do.

I can offer you no greater proof of that claim than to suggest you get an AI summary of Dame Edna's autobiography "My Gorgeous Life" and see if it makes you laugh.

I'm 93% certain it won't.