.png)
Most companies are busy with AI right now. A pilot is running in one corner, a few slides somewhere promise that a product is "AI-powered," and there may even be a budget set aside to go and write an AI strategy. What tends to be missing is a straight answer to the only question that counts: has any of it actually changed a result?
That gap, between the activity and the outcomes, is where most companies are sitting with AI. So we put it to three people running strategy and operations in very different industries: where is this technology actually earning its keep, and where is it just theater?
We ran a panel recently, co-hosted with The Strategy Brief, with Charlie Newark-French (CEO at Cascade), Joshua Carless (GM, Strategy at QBE), and Jamie Wiebe (Group Executive General Manager at Crown Resorts). None of them spent much time debating whether AI belongs in strategy, because all three are already using it. The honest conversation was about where it works, where it doesn't, and what stays human no matter how good the technology gets.
The 95% You Never See
Charlie likes to put one number in front of leadership teams. By the time information has climbed all the way up an organization, the people at the top are seeing about 5% of what's actually happening below them. And there's no guarantee it's even the right 5%.
Closing that gap is what AI is genuinely good at: it reads the other 95% and turns it into something close to a live picture of the business, fast enough to act on. Where Charlie stops short is letting it decide.
"AI is more of a thought partner. Gathering data, crunching the numbers. But accountability, human judgment, trade-offs, decision-making, that has to stay human when it comes to strategy. That's where companies differentiate themselves." - Charlie Newark-French
Charlie knows the work this replaces. Early in his career at McKinsey, a new project meant interviewing 60 people, asking the client for years of historic data, and slowly digesting it all to hand back what he'd heard. A model does most of that today, in a fraction of the time. What it can't do is the step that comes next: feed it everything and ask what to do, and you get a generic answer; ask it to pull the numbers and help you think, then make the call yourself, and you get something worth having.
Don't Lead With the Tool
Josh has a wide view across the function, through his network of CFOs and chief strategy officers, and he keeps seeing the same mistake: teams launch a pile of AI pilots, announce they're "embedding AI," and measure none of it. Busyness gets mistaken for progress. His advice is to go back to first principles.
"I take it back to the fundamentals of what strategy is, which is making long-term decisions to drive competitive advantage. I challenge organizations to challenge their own executive board on what the competitive advantage actually is, and you'd be surprised how misaligned some boards can be. Be really crisp on your points of differentiation. Why do you have loyal customers, why are you growing, why are you profitable? What's your edge? Once you know that, work out what capabilities you need to deliver value, and those capabilities can be AI or non-AI." - Josh Carless
Lead with the technology instead, in his words, and you've just got hammers looking for nails. Insurance is full of manual work and ageing systems, so he sees real opportunity, but the discipline is to chase the customer outcome and not get caught up in the theater of AI.
Jamie runs the same logic from inside one of the most scrutinized industries there is.
"AI is not the strategy, it's there to support the strategy. I like to separate the objective from the tool. At Crown we didn't set out to use AI for AI's sake. We had a problem: data that could help us identify risk earlier and prevent harms. But there's no way we could have an analyst going through hundreds of thousands of play indicators by hand. So we saw a real opportunity for AI to assist with our overall harm-reduction strategy." - Jamie Wiebe
Even then, the team turned down the tools that worked as a black box. They needed to know why a person had been flagged, because the next step is a real person having a careful conversation, and that conversation only helps if it's specific to the individual in front of you.
Plans Have to Move Faster Now
The more interesting question is what AI does to the shape of a plan.
Charlie framed it through an old distinction. A planned strategy picks a direction and commits to it for five years. An emergent one watches what's working and moves money toward it.
Starbucks in Australia is the cautionary tale of relying on a purely rigid, planned playbook. They rolled out roughly 300 stores in about two years using a textbook market dominance strategy. But Australians already took their coffee seriously, the independent café scene was strong, and the model fell flat. A few hundred million dollars later, they had to pivot.
The problem wasn't a lack of data. The baristas on the ground could have named the exact café beating them long before head office caught up. The problem was that it took Starbucks an incredibly long time to react to that data.
Because AI can give leadership a near-live read on the business, organizations can shift to an adaptable, emergent approach, allowing teams to hold the goal steady while reacting and tweaking execution as they go.
The Job Becomes Monitoring
If AI handles the analysis, the work left for people changes shape. Josh explained it through a friend who flies 777s for Emirates.
"At the gate he's programming the flight computer with the weight, the wind, the routes. Call that prompting. Once he takes off, it's largely a monitoring role. You need to understand what you're prompting and monitoring, and have the ability to intervene if it goes off the rails." - Josh Carless
That's a different skill from doing the work by hand, and it carries a quieter risk. A normal workday is a mix of simple, predictable tasks and harder problems. Take away all the simple work, and you leave people doing nothing but complex firefighting, all day, every day. That cognitive load is what tips into burnout.
It also breaks the way people learn. Charlie pointed out that twelve years ago, a McKinsey report predicted machine vision would end radiology within a decade. Instead, there are more radiologists now because the human parts of the job didn't disappear. But you only become a senior expert by doing the junior grunt work first, the exact work AI now absorbs.
Charlie noted that this structural talent gap is already highly visible in data from software engineering and call centers. If you are ten years into your career in those industries, your skills are more highly coveted and demanded than ever; fresh out of college, however, entry-level candidates are struggling to get hired because the basic tasks have been automated away.
"There's a lot of things you need to learn. Critical thinking, judgment, trade-offs, how to not shout back when someone shouts at you. We have to build those next jobs, and be part of designing the talent behind them." - Charlie Newark-French
The reassuring part is that none of this lands as fast as the headlines say. Charlie bet in 2014 that he could ride a driverless car anywhere in a city by 2020; it took until 2022, and only in San Francisco, where Waymo now runs more trips than Ubers or taxis. He once saw a bank use 50 separate systems to issue a single mortgage, one of them written back in 1970 in COBOL. Most organizations look more like that bank than a clean demo, so there's more time to get the people part right than the panic suggests.
Bring the Regulator With You
Adopting AI well isn't only an internal exercise. In an industry like Jamie's, where trust is the entire business, you can't quietly switch it on and hope nobody notices.
"Trust is critical to the sustainability of the business, and trust can take a very long time to build and seconds to disappear. So it's been really important for us to take the regulator on the journey with us, to be clear on what this technology is and is not, to moderate expectations, and make sure we're all on the same journey together." - Jamie Wiebe
Stewardship, the way she framed it, runs two ways. Inside the company it means real governance, with firewalls and safeguards around the data. Outside it means broader standards for how AI can and can't be used, because customers now arrive with raised expectations: if you hold their data, they expect it kept confidential and handled in their interest, and they can tell when it isn't.
What Stays Scarce Becomes the Premium
Near the end of the session, Josh delivered the line that stuck: I really like humans.
His case was a run of examples. The metaverse let you watch a film in a headset with friends, and people kept choosing the cinema. Amazon hollowed out the bookshops, yet the survivors are full of people who came for the experience. Dating apps are losing users to organic meetups like gyms and Parkruns. His sharpest example was his own industry: banks spent years pushing customers online, and now "come and talk to a real person at your branch" is a major source of market differentiation. The baseline became the premium.
That matters more as the next generation leans the other way. Around 70% of 18-to-25-year-olds say they'd trust AI over a human for financial advice. As the default shifts to the machine, human contact becomes the scarce thing, and scarce things carry a price.
Jamie has the numbers to back it. In the early days of Crown's risk work, a single human conversation cut measured risk by about half: one person reading one individual's situation and talking to them in a way that makes their wellbeing the obvious point. AI flagged who to talk to. The conversation did the rest.
"If you're not personalizing and not being responsive, you'll be left behind. The differentiator becomes how you use it, and what values are embedded in that." - Jamie
Charlie kept his own forecast to the window strategy leaders actually plan for, the next five to ten years. AI will run more of the analysis every year. What it can't do is choose a direction and own the call. That's still the job.
___________________________________________
This article is based on a panel discussion co-hosted by Cascade and The Strategy Brief, featuring Charlie Newark-French (CEO, Cascade), Josh Carless (GM, Strategy, QBE), and Jamie Wiebe (Group Executive General Manager, Crown Resorts), moderated by Devina Patel (Cascade). You can watch the full recording here.
.png)
.png)
.png)
