On execution, orchestration, and the leadership shift that AI is forcing faster than anyone expected.
The week I realized my biggest professional asset had become a liability, I was in a governance meeting where the team had reached an impasse — should we re-architect a legacy module in-house, or hand it off to the specialist team who could get it done? On the surface it was a resource and risk conversation. But it stopped being about the module pretty quickly.
I've spent the better part of my career building things — portfolios, programs, teams. I took pride in knowing how the pieces fit together: being the leader who could zoom into technical detail and then zoom back out to strategy. That fluency felt like my edge. But something shifted over the last 18 months, and the shift wasn't subtle.
AI tools are now doing the zooming-in. Not perfectly, not always reliably — but competently enough that the question is no longer "Can your team build this?" It's "Do you know what should be built, and can you get a team of humans and machines to build it well?" That's a different job. And it requires a different kind of confidence.
"The trap is doubling down on execution when the leverage has moved to judgment."
The trap I've seen smart engineers and product managers fall into is doubling down on execution when the leverage has moved to judgment. They become faster coders in an era that rewards better orchestrators. They optimize for output when the real premium is on outcomes. I get it — execution is comfortable. You can measure it. You get the dopamine hit of shipping. But when AI handles a significant chunk of the implementation layer, the person who wins is the one who is crystal clear on what problem is worth solving and why.
This hits hardest at the career inflection point tripping up a lot of people right now: the move from senior individual contributor to leader. I've talked to dozens of engineers and delivery professionals over the past year who are genuinely confused about where their leverage lies. They're exceptional builders, they understand the architecture deeply, and they're watching AI handle more of the tasks they were known for. Their instinct is to find harder technical problems — go deeper, stay specialized, outrun the automation.
That's one valid path. But there's another one that isn't getting enough airtime: becoming the person who can direct and evaluate AI output, and translate it into measurable business value.
I spent years thinking leadership was about having more answers. The most important thing I've unlearned in the last 18 months is that leadership is actually about asking better questions earlier in the process, and then creating the conditions for a team — increasingly a hybrid team of people and AI — to execute well.
A Deloitte study released earlier this year put it plainly: the mandate for tech leaders has shifted from "operators" to "orchestrators." That's not jargon — it describes a genuine change in where the value lives. The leaders generating the most impact right now aren't the best at executing; they're the best at context-setting, prioritizing ruthlessly, and building feedback loops that catch mistakes before they compound.
"The skills that make you indispensable aren't the ones that can be automated — they're the ones that make the automated work meaningful."
None of this means technical depth stops mattering. It doesn't. Judgment without grounding is just opinion. The leaders who will compound their value over the next decade are the ones who use their technical foundation to inform better strategy, not just better code. The governance meeting moment I described at the start wasn't resolved by deeper technical knowledge — it was resolved by clarity on business outcome, risk appetite, and team capability. That's the judgment AI can't replicate.
If I could tell my 10-years-ago self one thing, it would be this: learn to say "this is the right problem" before your team builds the wrong solution brilliantly. Taste. Judgment. Outcome ownership. These are the skills that compound — and they're the ones the shift to AI is making more valuable, not less.
When you think about the most important skill you've developed in the last two years —
is it keeping pace with the shift from execution to orchestration?
Or are you still optimizing for depth in a layer that AI is steadily climbing?
I'd love to hear where you land.