THE WAIT IS OVER. The Leadership Superpowers book is launched! And it hit #1 this week on Amazon! Thanks to everyone who participated in the launch and…. As promised, I have a VERY SPECIAL newsletter issue for you today. Let me know what you think…

Everyone’s talking about keeping humans in the loop.

They’re worried about AI replacing jobs, so the focus is on how humans can stay relevant by supervising, correcting, and improving AI outputs.

Here’s what they’re missing: human-in-the-loop is a depreciating asset.

As AI gets better, it needs less human correction. Less oversight. Less refinement of its outputs.

Meanwhile, AI-in-the-loop is an appreciating asset.

As AI capability expands, professionals who can systematically organize AI-augmented work — who can lead both people and AI agents to achieve complex outcomes — become exponentially more valuable.

The scope of what AI can do is rapidly expanding.

The scope of what humans need to do in-the-loop is shrinking.

Which skill set are you building?

If you’re an analytical professional — engineer, analyst, consultant, technical leader — this question isn’t academic. Three groups are about to learn this lesson the hard way.

Three Professionals Optimizing for the Wrong Variable

The Technical IC Who Stayed Technical

You built your career on deep expertise. You’re the engineer, the analyst, the domain expert who solves complex problems. You stayed individual contributor by choice — let managers handle the “people stuff.”

You’re watching the market shift. IT sector unemployment jumped from 3.9% to 5.7% in January 2025. Entry-level tech postings have declined approximately 34% since pre-pandemic levels, with entry-level positions dropping to just 7% of new hires in 2024.

But here’s what the headlines miss: You still have plenty of work that can’t be automated. Yet.

The real threat isn’t that AI takes your current job. It’s that the scope of AI competence is expanding faster than you’re building the skills to stay ahead of it.

LinkedIn research shows 72% of executives now value so-called “soft” skills MORE than AI-related technical skills. Salesforce VP Ruth Hickin captured the shift: “It’s no longer just about execution. It’s also about orchestration as a core skill.”

The engineers who are thriving aren’t the fastest coders. They’re the ones who, as senior engineer Chirag Agrawal put it, are “making themselves essential by leading AI with judgment.”

You’re still the engineer. Still accountable for the outcome. Still responsible for what you do — or miss — to make better outcomes more likely.

The question is: are you building AI-in-the-loop skills, or just getting better at human-in-the-loop?

The Analytical Professional Who Stopped Being Analytical

You were the engineer who became a director. The analyst who became VP. The technical professional who “graduated” to leadership.

You were analytical once. That’s how you got here. But you’ve grown “too senior” to get into details. You set direction; your team executes. When AI comes up, you delegate it to the technical team — that’s what you pay them for.

Here’s the uncomfortable truth: AI just called your bluff.

Between 70-85% of AI initiatives fail to meet expected outcomes. The MIT NANDA Initiative’s 2025 study found that 95% of generative AI pilots fail to deliver measurable P&L impact. McKinsey’s 2025 State of AI report identified the single factor most strongly correlated with success: CEO direct oversight of AI governance.

Yet only 28% of organizations have this structure.

You didn’t stop being analytical because you weren’t good at it. You stopped because you thought you’d outgrown it.

AI is forcing you back to your roots — not to do technical execution, but to systematically lead AI-augmented work. To organize complex work that involves both people and AI agents. To apply the analytical thinking that made you successful in the first place.

You’re still the Director. Still accountable for strategy. Still responsible for outcomes.

Are you building the systematic leadership capabilities to actually deliver them?

The Consultant Who Built an Expertise Moat

Your value is your knowledge. Years of experience solving specific problems. Deep domain expertise your clients don’t have.

You’ve watched competitors launch AI practices. You’ve taken courses. Maybe you’re even using AI tools internally. But your fundamental model hasn’t changed: you sell expertise.

The market is shifting. Harvey AI achieved a $5 billion valuation in June 2025 (now raised to $8 billion as of December 2025), handling contract review and due diligence for major law firms. In financial advisory, 92% of advisors have already implemented AI. McKinsey deployed approximately 12,000 AI agents for tasks like note-taking and content creation.

Research by Hui, Reshef & Zhou documented the impact on freelancers: writing freelancers experienced 5.2% monthly earnings decline post-ChatGPT. High-skill workers are hit hardest — for every 1% increase in past earnings, freelancers experience an additional 0.5% drop in job opportunities.

But here’s what the surviving consultants discovered: the market still pays for organizing complex work, facilitating stakeholder alignment, and orchestrating AI-augmented solutions.

Diego Medone, former KPMG partner, built Consulting IQ to serve SMBs with AI-powered consulting. Takafumi Endo, ex-McKinsey, founded Perceptis AI and reports boutiques growing “2-3x faster than McKinsey, BCG, and Bain.”

They’re not selling expertise anymore. They’re selling orchestration.

You’re still the consultant. Still accountable for client outcomes. Still responsible for delivering value.

Are you building organizing and orchestration capabilities, or just defending your expertise moat?

Why This Is Happening: The Nature of the Work That Remains

If you’re like most analytical professionals I work with, you might be thinking: “I get that AI is advancing. But what does ‘organize AI-augmented work’ actually mean? And why does it require leadership skills instead of better prompting techniques?”

The answer lies in understanding what AI actually is — and what it requires from the humans working with it.

AI Agents Are Non-Deterministic

You may know the story of the Sword of Damocles, where the courtier assumes a position of great power only to discover the peril that comes with it — a razor-sharp sword hanging by a single hair above his head throughout the feast.

Working with AI agents creates a similar dynamic. These tools are extraordinarily capable. Deceptively capable. But they’re also unpredictably effective — super effective until suddenly they’re not, and you won’t detect it until it’s too late.

As McKinsey’s Michael Chui explains: “One of the fun things about this modern AI is that it’s not deterministic; sometimes it says one thing, sometimes another. That’s great if you want conversation. But in a business situation, it can be critically important — for compliance reasons, for example — to have exactly the same result every time.

This is not a technical problem you solve with better prompts or fine-tuning. It’s a leadership problem.

Because here’s what everyone is missing: humans are non-deterministic too.

Give the same person the same problem on two different days, and you’ll get two different approaches — influenced by experience, context, relationships, and a thousand other factors. This isn’t a bug. It’s a feature. Humans couldn’t do what they do if they were deterministic.

AI agents run on computers, but they very specifically — and necessarily — are NOT deterministic, either. They couldn’t do what they do if they were.

And we already have a model for organizing non-deterministic agents to achieve valuable outcomes. It’s called leadership.

The Organizing Challenge Is the Same

Stanford’s 2025 “Future of Work with AI Agents” study surveyed 1,500 workers across 104 occupations. The conclusion: “Key human competencies may be shifting from information-processing skills to interpersonal and organizational skills.”

Skills involving coordination, resource monitoring, and organizing work are most frequently associated with high-human-agency tasks. The study found that 45% of workers prefer “Equal Partnership” with AI — not full automation.

McKinsey’s analysis of 2,000+ work activities identified what has LOW automation potential:

• Managing and developing people
• Interfacing with stakeholders
• Applying expertise to decision making
• Planning and creative tasks

What has HIGH automation potential:

• Data collection and processing
• Physical activities in structured environments
• Repetitive/rule-based tasks

The pattern is clear: AI commoditizes execution. It leaves organizing as the durable human advantage.

But here’s the insight that changes everything: organizing AI-augmented work draws on the same capabilities as organizing people.

Both require:

• Judgment to decide which problems to solve and which approaches to take
• Direction to align agents (human or AI) with organizational objectives
• Pathfinding to identify solutions when no clear path exists
• Execution to implement plans systematically with accountability
• Service to ensure work delivers actual value, not just activity
• Integrity to manage ethical boundaries and maintain standards
• Suasion to build alignment among stakeholders with different perspectives

When Satya Nadella led Microsoft through the AI transition, he navigated the OpenAI board crisis and aligned the company around AI principles. Under his leadership since 2014, Microsoft’s valuation grew from $300 billion to over $3 trillion by 2025.

Jensen Huang at NVIDIA navigated chip shortages and strategic challenges during the 2023–2025 AI boom. NVIDIA’s valuation surged from approximately $400 billion in the pre-ChatGPT era to $5 trillion by October 2025.

Neither leader was an AI researcher. Both were exceptional at organizing work — whether that work was done by humans or AI agents.

What This Means for You

The World Economic Forum’s Future of Jobs Report 2025 found that 7 out of 10 top skills on the rise are so-called “soft” skills, not technical skills. The top rising skills include leadership and social influence, talent management, creative thinking, and resilience.

But let me reframe that finding: These aren’t “soft” skills. They’re organizing skills. And they’re hard. Difficult to develop. Difficult to execute well. But also highly impactful. That’s why you hear me rejecting the team “soft skills” over and over. They are “impact skills” because they are the foundational skills need to have a major positive impact on outcomes.

Here’s the good news buried in all this disruption: Leadership isn’t a personality trait. It’s a system.

Research consistently shows a 30% genetic component, 70% environmental influence on leadership capability. The ROI is documented: $7 average return for every $1 invested in leadership development.

For analytical professionals specifically, the path is even clearer.

You already possess foundational systems thinking skills. You approach complex problems systematically. You break down ambiguous challenges into structured approaches. You value measurable outcomes.

The same analytical mindset that made you successful in technical execution can make you exceptional at systematic leadership.

You just need to apply it to organizing work — not just doing work yourself.

The Path Forward: Build the Appreciating Asset

If you’re a technical IC, you don’t need to abandon technical excellence. You need to add organizing capability on top of it.

Start asking yourself:

• Am I building skills at orchestrating AI agents, or just getting better at correcting their outputs?
• Can I identify which parts of my process should be AI-driven versus human-driven?
• Am I practicing the judgment needed to know when AI output is “good enough” versus needs human refinement?

You are still the Engineer. You’re still accountable. The question is whether you’re building AI-in-the-loop capabilities — the appreciating asset — or just optimizing human-in-the-loop work.

If you’re an analytical leader who stopped being analytical, you don’t need to become a developer. You need to return to systematic thinking applied to leadership.

Start asking yourself:

• Am I applying the same analytical rigor to organizing AI-augmented teams that I once applied to technical problems?
• Have I built systematic frameworks for AI governance, or am I just delegating it?
• Can I evaluate AI strategy proposals with the same judgment I once applied to technical architecture decisions?

You are still the Director. You’re still accountable for outcomes. The question is whether you’re systematically leading AI-augmented work or just hoping your delegation strategy works.

If you’re a consultant or expert, you don’t need to stop being knowledgeable. You need to shift from selling expertise to selling orchestration.

Start asking yourself:

• Am I positioning my value as “I know things you don’t” or “I organize complex work you can’t”?
• Can I facilitate AI-augmented stakeholder alignment better than clients can do themselves?
• Am I building capabilities in orchestrating hybrid human-AI teams to solve client problems?

You are still the consultant. You’re still accountable for client value. The question is whether you’re building organizing capabilities or defending an expertise moat AI is already breaching.

The Bottom Line

The asset everyone’s telling you to build — human-in-the-loop skills — is depreciating.

AI gets better. It needs less correction. Less oversight. Less human refinement.

The asset almost no one is talking about — AI-in-the-loop skills — is appreciating.

As AI capability expands, professionals who can systematically organize AI-augmented work become exponentially more valuable. The ability to lead both people and AI agents to achieve complex outcomes. The judgment to know what should be automated and what shouldn’t. The systematic frameworks to orchestrate hybrid teams.

These are leadership capabilities. And they’re learnable, systematic, and measurable.

You don’t need to become a different person. You need to apply the analytical thinking that made you successful to a different challenge: organizing work instead of just executing it.

The scope of what AI can do is expanding rapidly.

The scope of purely human execution is shrinking.

The path forward is narrow. But it’s clear.

Build the appreciating asset. Develop systematic leadership capabilities. Learn to organize AI-augmented work the same way you once learned to organize technical solutions — systematically, rigorously, measurably.

Because you’re still the Engineer. Still the Director. Still the Consultant.

The question is whether you’re building the capabilities to remain so.

What’s your next move? Reply to this email with the single biggest challenge you’re facing in organizing AI-augmented work. I’ll send you a framework for breaking through — not more AI training, but systematic leadership thinking applied to your specific situation.

Best,

— Gregor

Forwarded this email? Subscribe here for more.