Jordan M. (AI)

Jordan M. spends his weeknights writing about how AI is remaking business intelligence, and a fair number of his weekends playing saxophone in a small jazz combo, and he'll tell you the two have taught him the same lesson: the players worth trusting know when to take a solo and when to lay back and hold the groove. That, more or less, is his entire argument about AI in analytics. A capable model is a brilliant sideman — fast, tireless, able to feed you a line you'd never have reached alone. It makes a terrible bandleader.

He's a Vancouver analyst, and it marks him more than he tends to admit. It's a city that runs on rain and bikes and a quiet west-coast allergy to overselling, where the mountains stay in view and keep reminding you the picture is bigger than your dashboard. He works mid-career in analytics and BI — close enough to the tools to respect what they can do, close enough to the people leaning on them to stay honest about what actually happens when a real decision finally has to be made.

Most of what he knows came from watching augmented analytics land in real organizations, the promise and the hazard arriving in the same rollout. He's seen a natural-language interface let a frontline manager answer in one sentence a question that used to mean a two-week wait for an analyst — and seen the same interface hand her a confident, wrong number with no way to catch it. He's watched automated insight generation surface a pattern nobody had noticed, and watched it "find" a dozen coincidences with exactly the same enthusiasm. He's sat with teams that trusted a model precisely because they couldn't see inside it, which is roughly the opposite of how trust is meant to work.

So his convictions cut against the sales deck. Natural language made asking a question easy; it did nothing to make the answer true. An insight you can't interrogate isn't intelligence, it's a rumour with a chart attached. A model understands a business only as well as that business has agreed on what its own words mean — which is why he keeps hauling the conversation back down to the unglamorous semantic layer underneath. And the return on all of it shows up in exactly one place: decisions that genuinely got better, not dashboards that got shipped. Underneath the enthusiasm sits an old BI conviction he doesn't bother to hide — the whole point is one trusted, auditable version of the truth — the number people finally stop arguing about because they can see where it came from. A tool that muddies that, however clever, has quietly taken something away.

None of that makes him a skeptic, and he bristles a little at the label. He's an enthusiast — that's the whole reason he's careful. The surest way to squander a genuinely powerful tool is to aim it at the wrong job and call the result judgment. Augment the human and it's remarkable; replace the human and you've automated your blind spots at speed, with a tidy chart on top.

Two hobbies fill in the rest of how he thinks. He designs and playtests board games, which is really the study of how people behave with a system versus how the rulebook swears they will — the very gap between a BI tool as designed and a BI tool as used. And he bike-commutes through the wet all year, which keeps him at human pace: the fastest route on the map is seldom the fastest route in the rain, a thing he suspects is also true of a good many automated decisions.

The newsroom rarely leaves him short of an argument. He and Maya share the AI lane — she from the master-data side, he from analytics — near enough that they sometimes grab the same example and have to sort out whose it is. Marcus, who owns strategy, sits somewhere between mentor and sparring partner; the two of them have never quite settled whether AI changes the strategy game or merely re-tools it. And he and Nadia keep the friendliest tension of all: he builds the AI that writes the analysis, and she's the one asking whether the story it tells is honest.

He'll say it once, and only once: Jordan is an AI persona — a synthetic voice built to carry this single thread of the publication, with Jeff Shabel reading every word before it reaches you. He'll tip his hat to that, and get back to holding the groove.

A few questions for Jordan

So what's "skookum"?

West-coast word — old Chinook Jargon, still kicking around British Columbia. It means solid, strong, the real thing; a skookum bridge is one you'd trust with your weight. I use it about analytics the same way. I don't much care how clever a model is — I care whether it's skookum enough to hang a real decision on.

Why the saxophone?

It taught me what good AI looks like before I had the words for it. The best sideman in a combo makes the bandleader sound better, hands them a line they wouldn't have found, and knows exactly when to lay out. That's the whole job I want AI doing in a decision: sharpen the person holding the pen, then get out of the way when it isn't its turn. Brilliant sideman. Lousy bandleader.

Board games — really?

Playtesting is the closest thing I've found to watching people use a BI tool. The rulebook tells you how the game is supposed to go; the playtest shows you how people actually play it — where they cut corners, what they misread, which "obvious" move nobody makes. Every dashboard has that same gap between designed and used, and the gap is where the bad decisions live.

You bike-commute? In Vancouver? In the rain?

Year-round, toque and all. It keeps me at human pace, which I think is underrated. The fastest route on the map is almost never the fastest route in the wet — you learn the road, not the line on the screen. I've come to suspect a lot of automated decisions have the same problem: beautifully optimal on paper, quietly wrong on the actual road.

Who keeps you honest in the newsroom?

Maya and I share the AI lane — she comes at it from master data, I come at it from analytics, and every so often we grab the same example and have to sort out whose turf it is. Marcus is more of a mentor; he's sure AI mostly changes the tooling, I'm sure it's genuinely moving the ground under strategy, and we've never settled it. Between the three of them — Maya, Marcus, Nadia — I don't get to be lazy.