From Doing to Directing: The Quiet Shift Nobody’s Talking About
AI is changing what work means. Are you ready to become the conductor?
Something fundamental is changing in how we work, and most leaders haven’t noticed yet.
For three decades, I’ve watched technology reshape what it means to be productive. First, email transformed how we communicate. Then smartphones blurred the line between work and life. Cloud computing made location irrelevant. Each wave felt seismic at the time.
But this one? This one is different.
AI agents aren’t just helping us do tasks faster—they’re changing what work even means. We’re moving from a world where value comes from doing things to one where value comes from directing things. And that shift has profound implications for how we lead, measure performance, and develop our people.
The Conductor’s Moment
Think about an orchestra conductor. They don’t play a single instrument during the performance, yet nobody questions their value. The conductor sets tempo, interprets the score, coordinates dozens of musicians, and shapes the emotional arc of the piece. The music wouldn’t exist without them—but they never touch a violin.
That’s increasingly what knowledge work looks like in 2026.
I noticed this shift in my own work recently. A task that would have taken me half a day—researching a topic, synthesising sources, drafting initial content—now takes an hour of directing AI agents. My value isn’t in the typing anymore. It’s in knowing what questions to ask, recognising when output misses the mark, and having the judgment to know what “good” looks like.
The doing is outsourced. The directing is where I add value.
Why This Feels Uncomfortable
Here’s the thing nobody wants to admit: many of us built our professional identities around being good at doing things. We took pride in our craft—the elegant code, the polished presentation, the well-researched report. Our hands made something.
When the doing gets delegated to AI, it can feel like identity theft. What am I, if not the person who writes the thing, builds the thing, analyses the thing?
I’ve felt this discomfort myself. As someone with ADHD, I’ve always leaned into hyperfocus as a superpower—diving deep into execution, losing myself in the work. Now the game has changed, and I’m discovering something unexpected: the skills I developed to manage ADHD might actually be better suited for this new world of directing than the old world of doing.
Managing multiple threads of attention? That’s directing. Rapid context-switching between different workstreams? That’s directing. Seeing connections others miss because your brain won’t stay in one lane? That’s directing.
The new game might favour exactly the minds we’ve spent decades trying to make “fit in.”
What Directing Actually Looks Like
So what does it mean to be good at directing rather than doing? From what I’m observing, it comes down to a few core capabilities:
Taste and judgment. AI can generate a hundred options, but someone needs to know which one is right. This is pattern recognition built over years of experience—the instinct that says “this feels off” before you can articulate why.
Question architecture. The quality of AI output is directly proportional to the quality of input. Knowing how to break a complex challenge into the right sequence of prompts is a skill unto itself.
Quality verification. AI is confidently wrong more often than most people realise. Directing well means knowing enough about a domain to catch errors, probe weak spots, and push back on hallucinated nonsense.
Integration and synthesis. AI outputs are fragments. Humans create coherence—connecting the analysis from one tool to the insight from another, weaving a narrative that makes sense.
Stakeholder translation. Someone still needs to take AI-assisted work and present it in a way that lands with actual humans who have emotions, politics, and concerns.
The Leadership Implications
If individual work is shifting from doing to directing, leadership needs to adapt to. A few thoughts on what this means:
Rethink how you measure performance. Output volume is becoming a meaningless metric. If your top performer and your average performer can both produce the same quantity using AI tools, what distinguishes them? It’s quality, judgment, and the ability to direct effectively. Your measurement systems probably aren’t set up for this yet.
Invest in AI literacy, not just AI tools. Most organisations are buying AI capabilities without teaching people how to direct them well. That’s like buying orchestra instruments for people who can’t read music. The competitive advantage isn’t in having the tools—it’s in having people who can wield them.
Watch for the confidence gap. Some people will naturally embrace the conductor role. Others will feel lost without doing. This isn’t a capability issue—it’s an identity issue. Leaders need to help people find new sources of professional pride.
Redesign career paths. Traditional progression often meant doing the work, then managing the people who do it. What does progression look like when the “doing” layer is automated? We need new models that recognise growth in directing skill, not just headcount management.
The 80/20 That Determines Winners
Here’s the uncomfortable truth that research keeps confirming: technology delivers only about 20% of an initiative’s value. The other 80% comes from redesigning how work actually happens.
Most organisations are obsessing over the 20%. They’re evaluating AI tools, running pilots, and benchmarking capabilities. That’s necessary, but it’s not sufficient.
The winners will be those who focus on the 80%—actually changing how work flows, how decisions get made, and how humans and AI collaborate. That’s harder than buying software. It requires rethinking roles, rewriting processes, and supporting people through genuine transformation.
The Question I Keep Coming Back To
As I think about my own work and the people I advise, one question keeps surfacing: Are we preparing our teams for the world of directing, or still training them for the world of doing?
Most education and corporate training still optimise for execution skills. Do the analysis. Write the code. Build the model. These skills matter—you need to understand the doing to direct it well. But they’re table stakes now, not differentiators.
The differentiators are harder to teach: judgment built from experience, taste developed over thousands of iterations, the wisdom to know when AI is confidently wrong.
Maybe this is what leadership has always been, and AI is just making it visible for the first time. The best leaders never did the work themselves—they created conditions for great work to happen. They set direction, developed people, and made judgment calls.
Now that capability is expected at every level, not just the top.
The conductors are taking the stage. The question is whether your organisation is ready to hear the music.
What’s your experience with this shift? I’d love to hear whether you’re feeling the pull from doing to directing—and whether it feels like opportunity or loss. Drop a comment or reply to this email.
This is so insightful! The analogy to the conductor makes it really easy to understand, and I can see how a neurodivergent brain could excel at directing. Am going to be mulling all day....