Episode transcript
Note: This transcript has been edited for clarity and readability while staying true to the spirit of the conversation.
Hello and welcome to Judgement Calls: Leading in the AI era.
I’m Carl Raath, and this is the weekly podcast for leaders who want to elevate their own impact and the impact of the people they lead.
Here’s this week’s idea.
There’s a message I keep hearing in leadership conversations, and every time I hear it, I wince a little. The sentence is something along the lines of, “We’ve done AI.”
Sometimes it arrives with a slide: licences rolled out, training done, a usage chart heading up and to the right. Sometimes it’s just a tone in the room—a sense that this one’s been handled and we can move on to the next agenda item.
And look, I get where it comes from. The visible work was genuinely done. Tools were bought. People were shown how to log in. Something was announced at a town hall.
But my view is that “we’ve done AI” might be the most expensive message in business right now.
The rollout itself was probably fine. The damage is in what that sentence does afterwards. Once a leadership team believes the box is ticked, nobody goes looking for what was actually supposed to change. And with AI, nearly everything that matters happens after the rollout.
I find one distinction does most of the work here.
Giving everyone a Copilot or ChatGPT licence is access.
Changed ways of working—things you can point to and measure—that’s adoption.
New capability—things the organisation can do now that it simply could not do before—that’s transformation.
Three different things. But the expensive mistake is buying the first one and quietly telling yourself you’ve got the third.
Nobody ever stood up in a board meeting and said, “We’ve done the internet.” It sounds ridiculous because we all understand the internet as a capability you keep building on, rather than a project you can close out.
AI has the same shape. There’s no fixed line beyond which you can declare yourself finished. Organisations treating it as procurement with an end date are going to keep getting surprised.
The numbers on this are fairly blunt.
McKinsey found that 92% of companies plan to increase their AI investment over the next three years, while only around 1% describe themselves as mature in how they use it.
So nearly everyone is spending the money and almost nobody has finished the work. That means an awful lot of leadership teams are somewhere in the middle, holding a dashboard that says “access” and a belief that says “done”.
PwC’s CEO survey tells a similar story from another angle. Most CEOs report little or no revenue or cost benefit from AI yet. But a small group—roughly one in eight—are seeing both.
When you look at what separates that group, everyone has access to similar tools now. So it can’t just be the tools. What the frontrunners did differently was change how the work actually gets done—and treat that as a leadership job, rather than something to hand to IT.
That last part matters more than most executives realise.
Microsoft’s workplace research found that organisational factors such as culture, manager support and talent practices account for more than twice the reported impact of individual factors such as mindset and behaviour.
So, if adoption is lagging in your business, I’d resist the urge to blame your people’s prompting. I’d look at whether the leadership team has made working differently safe, expected and rewarded.
That question lands on your desk.
There’s also a piece of MIT research that received a lot of attention recently: the finding that about 95% of enterprise generative AI pilots showed no measurable P&L return.
I’m really careful with that one because it’s preliminary work and it was measuring hard P&L impact within about six months, which is a brutal bar.
Reading it as “AI doesn’t work” would be the wrong lesson.
Read properly, it says something more specific: access without changed ways of working produces pilots that go nowhere. The technology holds up fine. Stopping at the rollout is what kills it.
For those of us in Australia, there’s a local wrinkle.
Deloitte’s latest global survey found that around 65% of Australian organisations intend to increase their AI investment, compared with roughly 84% globally.
We’re a step behind on the numbers, which cuts both ways.
If you’re benchmarking yourself against your neighbours and feeling comfortable, that’s a risk. If you’re prepared to actually do the adoption work while your local competitors are still ticking boxes, that’s a genuine opportunity.
The pattern I keep seeing has the same shape every time.
Licences go out. Usage spikes for a month and then settles into a small band of enthusiasts. The board gets a slide saying AI was adopted.
Meanwhile, the people closest to the work are quietly further ahead than the executive team assumes, using the tools in ways nobody sanctioned and nobody is learning from.
So, the formal program has stalled and the informal one is invisible.
The belief that AI is done is exactly what’s blocking progress. A solved problem attracts no attention, no budget for the hard part and no curiosity.
So, how do you know which side of the line you’re on?
Here’s a simple test: would your team say its decisions are measurably better because of AI?
Faster emails and tidier documents don’t count. I mean better decisions, better problems chosen and better questions asked in the room.
If you’ve got access everywhere and you can’t say yes to that question, then access is what you have. Adoption hasn’t started—whatever the dashboard says.
So, here’s what I’d do this week, and it costs you one meeting.
Take your most recent AI update—whatever you sent to the board or executive team—and put one question to it.
Skip how many people are using it and ask: what has changed about how we work? And what can we measure now that we couldn’t before?
If the room can genuinely answer, you’re further ahead than most.
If the answer turns out to be a usage statistic, you’ve learned something important—and you’ve learned it cheaply.
The sentence worth aiming for is, “Here’s what we can do now that we couldn’t a year ago.”
“We’ve done AI” closes the conversation.
The other sentence opens up an actual advantage—and the only person who can keep that search open is you.
That’s it for this week. Thank you for listening or watching, and for being the kind of authentic leader people choose to follow.
I’ll see you next week.
