Resourcefulness vs Agency: What We Mean When We Say "Agentic"

“We’re building agentic AI.” “I’m looking for more agentic employees.” “Our company needs to be more agentic in this market.”

The word “agentic” has entered the lexicon as one of those terms everyone uses but few can define precisely. We recognize it when we see it, but ask people to explain exactly what makes something “agentic” and you’ll get the hand-waving equivalent of someone trying to explain how a can opener works without ever having examined one closely.

Here’s a revealing test: Hand someone a can but no can opener. The resourceful person will google solutions, find tools to substitute, or develop clever workarounds within the constraints of “must open this can with what’s available.” The agentic person first questions whether they need what’s in the can at all, and — if they do — smashes the can with the nearest available blunt object.

Both approaches have merit. But they represent fundamentally different modes of intelligence that we conflate at our peril.

The Player and the Game Designer

Concrete definitions cut through the fog:

Resourcefulness is optimization within constraints. It’s the middle manager improving process efficiency by 15%, the engineer reducing code complexity while maintaining functionality, the marketer A/B testing to boost conversion rates. It accepts the system as given and seeks advantage within its rules.

Agency is constraint transformation. It’s the executive who realizes the company is in the wrong business entirely, the developer who eliminates an entire technology layer rather than optimizing it, the strategist who abandons a market instead of competing better within it. It questions which rules should exist at all.

The distinction isn’t abstract—it appears in measurable outcomes:

The resourceful product team adds requested features faster and with fewer bugs
The agentic product team eliminates the need for those features by solving the underlying problem differently
The resourceful CEO improves profit margins within the existing business model
The agentic CEO pivots to an entirely different business model with fundamentally different economics

Your organization claims to want “agentic” people but its reward systems, metrics, and culture probably optimize for resourcefulness. This isn’t hypocrisy—it’s the natural tension between executing today’s business and creating tomorrow’s.

I should acknowledge something important here: most people aren’t particularly strong in either dimension. The world is full of individuals who neither optimize brilliantly within systems nor transform systems themselves. They simply participate adequately, collecting paychecks, meeting expectations, and living lives untroubled by the burning desire to either master or reinvent their environments.

And that’s entirely fine. Not everyone needs to be wildly successful in taking markets through resourcefulness or making markets through agency. Not everyone needs to climb corporate ladders or start companies.

In fact, I’ll admit my own limitation here—I’m probably not spiky enough in either direction. Like many, I fancy myself as having flashes of agency but often lack the relentless resourcefulness to execute fully on those visions. The truly impressive people I’ve met tend to be extreme in at least one of these dimensions, and the most remarkable are those rare individuals with the capacity for both.

This distinction isn’t merely philosophical—it has profound implications for how we build technological systems and organizations, as I’ll explore in a companion piece.

The Anatomy of Agency: Strange Loops and Meta-Models

Most people conflate agency with creativity or vision, but that’s a category error. True agency requires a specific cognitive architecture—a way of processing reality that goes beyond mere imagination.

At its core, an agent is a strange loop—a system that models itself modeling a world. This recursive structure isn’t merely philosophical; it’s the computational foundation of agency itself. A thermostat responds to inputs but doesn’t model itself. A true agent maintains a self-model that it can examine and modify, creating the cognitive foundation for genuine transformation.

Agency emerges from three specific capabilities that most people and organizations lack:

The ability to see constraints as variables rather than constants. McKinsey consultants see process inefficiencies; agency sees that the entire process might be unnecessary. Your resourceful CFO optimizes the budget; an agentic one questions whether the accounting paradigm itself makes sense for your business.
Self-reflexive attention that allows for modification of one’s own thinking patterns. Resourceful people optimize within their mental models; agentic people modify the models themselves. This is why truly agentic thinkers often seem strange—they’re operating on metamodels invisible to others. When Elon Musk questions whether tunnels must be expensive or Jeff Bezos reimagines retail, they’re exercising this self-reflexive capability.
Recursive goal structures that enable questioning of objectives, not just paths to achieve them. If you’ve ever worked with someone who solved what seemed like the wrong problem but turned out to be right, you’ve witnessed this in action. This recursion creates a hierarchy of agency—first-order agents operate within models of how the world works, second-order agents modify those models, and third-order agents modify their modeling processes themselves.

This is why “innovation theater” fails so predictably in large organizations. They staff innovation labs with resourceful people, then ask them to exercise agency—a category mismatch doomed from the start.

But there’s a deeper truth: even when they claim to want agency, large organizations are fundamentally structured to reject it. This isn’t hypocrisy—it’s existential self-preservation.

True agency threatens everything that makes a large organization function: its hierarchy, its business model, its investment thesis, its quarterly projections. The CEO who publicly demands “more innovation” and “disruptive thinking” is simultaneously accountable to a board that demands predictable quarterly results and risk mitigation.

So organizations develop sophisticated immune responses to agency:

They create “innovation labs” physically and organizationally isolated from core business
They promote “intrapreneurship” programs with carefully managed boundaries
They celebrate “moonshots” that never threaten existing revenue streams
They hire “change agents” who lack actual authority to change anything material

The result is a cruel paradox for the genuinely agentic person within these environments. They’re told to “think differently,” but when they actually do—when they question fundamental assumptions about the business, challenge sacred cows, or propose truly disruptive directions—they encounter resistance proportional to how genuinely agentic their thinking is.

This isn’t because large organizations are malicious or dishonest. It’s because they’re optimized for preservation and execution, not transformation. They’re machines built to execute today’s business model, not to question whether that model should exist.

When Agency Threatens Organizational Survival

The previous observations reveal something more profound about why large organizations struggle with agency. When a middle manager suggests that the company’s core product category might be obsolete, they’re not just offering an opinion—they’re threatening the organization’s very existence.

This isn’t perceived threat—it’s actual threat. If the middle manager’s analysis is correct, and the core business is truly obsolete, then the rational response might be to dismantle large portions of the organization. But organizations, like organisms, are wired for self-preservation.

The organizational response is entirely predictable:

The agentic middle manager doesn’t get promoted—they get sidelined
Their insights are labeled “not practical” or “not aligned with our strategy”
They’re advised to “be more of a team player” and “focus on execution”
Eventually, they either conform or leave

Meanwhile, the resourceful manager who delivers 12% better efficiency in the core process (even if that process is becoming obsolete) receives bonuses and promotions. The system knows what ensures its continued existence, and it isn’t agency.

This dynamic creates the “disrupt or be disrupted” cycle that defines modern business. Organizations optimize for resourcefulness until they’re blindsided by challengers operating with agency. After the disruption, survivors institutionalize the new paradigm—converting yesterday’s agency into today’s resourcefulness, starting the cycle anew.

The most honest organizations acknowledge this tension explicitly rather than pretending to want something they’re structurally incapable of rewarding.

This explains the curious case of roles with “transformation” in their titles. The Director of Finance Transformation, the VP of Digital Transformation, the Chief Transformation Officer—these impressive-sounding roles almost never involve actual transformation.

Instead, they typically mean “optimizing what we already do”:

“Finance Transformation” means implementing better ERP systems, not questioning whether GAAP accounting makes sense for digital business
“Digital Transformation” means moving existing processes to the cloud, not fundamentally rethinking work in a digital paradigm
“Supply Chain Transformation” means incrementally improving logistics, not creating radically different supply models

The transformation these roles deliver is resourcefulness masquerading as agency. They improve the efficiency of existing processes by 25%, reduce costs by 18%, increase speed by 40%—all valuable outcomes, but all within existing paradigms. Real transformation—questioning whether the finance function should exist in its current form, whether digital requires entirely new business models, whether the supply chain concept itself is outdated—remains firmly off-limits.

This isn’t a criticism of the people in these roles. They’re playing the game as defined. The problem is semantic dishonesty—using “transformation” to describe what is actually “optimization.”

The Resourcefulness-Agency Paradox

These dynamics reveal several paradoxes familiar to anyone who’s tried innovating within established systems:

The Constraint Paradox: Resourcefulness thrives under constraints while agency requires their absence. This is why startups can reinvent industries while incumbents can’t reinvent themselves—different constraint environments produce different cognitive modes.
The Expertise Trap: Deep domain expertise often impedes agency. After you’ve spent 10,000 hours mastering a system’s rules, questioning those rules feels like abandoning your investment. The best consultants make terrible founders precisely because they’ve optimized for resourcefulness.
The Implementation Gap: Agency without resourcefulness produces compelling visions that fail in execution. This explains the visionary CEO whose brilliant strategies never quite materialize—all agency, insufficient resourcefulness.

These aren’t just abstract observations. They’re predictive patterns that explain why certain organizations consistently produce incremental improvements but never breakthroughs.

Cultivating Both Modes: Meta-Optimization and Recursive Intelligence

The most successful individuals and organizations develop meta-cognitive capacity—the ability to switch deliberately between resourcefulness and agency depending on context. This isn’t just flexibility; it’s a higher-order cognitive function—optimizing the optimization process itself.

This recursive capability explains why certain rare executives succeed across multiple industries. They achieve a strange loop in their cognition—modeling not just the business environment but their own modeling process. They can toggle between cognitive modes that others use exclusively, playing the existing game brilliantly when appropriate and designing new games when necessary.

What we’re observing in these exceptional cases is a hierarchy of intelligence:

First-order intelligence: Resourcefully solving problems within given constraints
Second-order intelligence: Questioning and modifying the constraints themselves
Third-order intelligence: Developing systems that know when to apply first or second-order approaches

In practice, this meta-optimization means:

Segregating resourcefulness and agency organizationally (operational teams vs. exploration teams)
Cultivating temporal boundaries (operational periods vs. strategic retreats)
Developing leadership that can recognize which mode is appropriate when
Creating feedback loops that question not just performance but the metrics of performance themselves

The organizations that master this recursive loop don’t just outperform—they outlast. They extract maximum value from existing paradigms while simultaneously pioneering new ones, creating the kind of exponential improvements we associate with both technological progress and biological evolution.

The Future Belongs to Strange Loops

The AI revolution brings this distinction into sharp relief. Current AI systems embody pure resourcefulness—they optimize within constraints at superhuman levels but cannot yet question the fundamental constraints themselves. They lack the self-reflexive modeling required for true agency.

This creates a profound inflection point. As machines monopolize optimization within constraints, human value will increasingly derive from constraint transformation—from agency. The future doesn’t belong to those who compete with AI at playing the game better, but to those who can identify which game is worth playing at all.

This shift elevates the importance of developing what Hofstadter called “strange loops”—cognitive structures that can reflect on and modify themselves. True agency requires this recursive capability—the ability to not just model the world but to model ourselves modeling the world.

When we describe someone as being “agentic,” we’re actually pointing to this self-recursive quality—their ability to question not just the answers but the questions themselves, to modify not just the solution space but the problem definition.

Perhaps what we call wisdom is precisely this self-modifying dance—knowing when to deploy recursive thinking, when to accept constraints and when to transform them, when to play the existing game brilliantly and when to create entirely new games.

The most valuable organizations and careers of the coming decade will be built on cultivating these strange loops. They’ll thrive not by choosing between resourcefulness and agency, but by mastering their recursive interplay and developing the meta-cognitive capacity to know which mode serves best in which context.