The Agency Advantage: Making AI Work Requires Human Judgment

Part 1 – The Agency Advantage is a three-part series exploring one of the most underrated drivers of organisational performance: agency. Whether you are leading a team through change or designing the conditions for your people to thrive, this series will take you through what agency is, how it shows up at work, and what leaders can do to strengthen it.

Your Teams Have the “AI Enablers”. So Why Isn’t It Working?

Singapore’s digital economy grew by S$12 billion in 2024, reaching 18.6% of GDP. AI investment is accelerating, and most organisations have already deployed tools, run pilots, or mandated adoption. You may have sat through the demos, signed off on the licences, and sent the all-staff email. Results, for most, have been underwhelming. That disparity comes down to a quality called agency, and building it may be the most practical thing leaders can focus on right now.

84% of organisations report adopting AI globally

11% actually realise meaningful value from it

74% of SG workers use AI, yet impact stays inconsistent

Sources: McKinsey & Company, 2025; IMDA Singapore, 2025

Agency is described as people’s belief in their capacity to act in ways that shape outcomes, rather than simply reacting to circumstances. In the workplace, agency is the degree to which individuals take deliberate ownership of decisions, adapt how they work, and act on their own judgement rather than waiting for direction. Some people have a stronger sense of it than others, and the same person can show more or less of it depending on the environment they are in.

With Agency vs Without Agency

Think about the colleague you have worked with who pushed back on a bad process, flagged a problem before it became a crisis, or kept a project moving when no one was sure what the next step was. Now think about the ones who waited for the memo, avoided the decision, or said "that's above my pay grade" when something needed to be resolved.

While agency is often viewed as a fixed trait, it is actually a dynamic interaction between individual capacity and the environment. To understand how these forces align, we adopt the following definitions: 

• Personal Agency (PA): An individual’s capacity to act intentionally, make choices, and influence outcomes despite constraints. “I can act and influence what happens to me and around me.”
  

• Organisational Agency (OA): The organisation’s collective ability to intentionally act, adapt, and shape its environment, enabled by leadership, culture, structures, and systems. “We, as a system, can act and influence outcomes.”

The Agency Matrix

When these two forces are mapped together, four distinct organisational behaviours emerge:

	Low Organisational Agency (Restrictive Culture)	High Organisational Agency (Empowering Culture)
Higher Personal Agency	Frustrated Drivers:  These employees push back on bad processes and try to solve problems, but are often blocked by bureaucracy or hierarchy, leading to high burnout and attrition risk.	Empowered Owners:  The "sweet spot" for transformation. These individuals challenge broken processes, raise issues early, and maintain momentum even in ambiguity.
Lower Personal Agency	Constrained Passenger:  In toxic or overly controlled environments, individuals default to compliance over contribution and silence over risk-taking to stay safe.	Enabled Passivity:  Even when leadership provides tools and empowerment, individuals may still stall, avoid decisions, or wait for explicit training before attempting anything new.

To shift a workforce toward the "Empowered Owner" quadrant, organisations must address the key drivers of agency:

• Psychological Safety: High agency cannot exist in a culture of fear. People must feel safe to take risks and offer constructive dissent.

  
  

• Purpose and Outcome-Based Expectations: Agency flourishes when people are given a clear “why” and "what" (the reason for change and goal) rather than a rigid "how" (the task). Over-indexing on processes trains employees to stop thinking and start executing within a box.

  
  

• The "Multiplier" Effect of Agency: Intelligence and talent are often static, but agency acts as a multiplier. Someone with slightly less raw intelligence but high personal agency will often create more value because they adapt to their environment to create leverage.

  
  

However, this multiplier effect works both ways. Just as high personal agency can accelerate growth, a lack of it creates a passivity gap that stalls progress and devalues even the most talented workforce. Whether it manifests as the stagnation of a "Constrained Bystander" or the under-utilised potential of "Enabled Passivity," the result is a massive drain on organisational energy. Gallup research finds that this lack of active engagement costs companies up to $550 billion annually in lost productivity. 

This discrepancy is becoming most visible in the rollout of generative AI. Because AI requires users to constantly exercise judgment and work their way through new workflows rather than follow a set manual, agency has become the key multiplier of talent.

Agency and AI-Enabled Work

As routine tasks get automated, with McKinsey estimating that up to 60–70% of current work activities have automation potential, the remaining “workload” demands interpretation, judgment, and decision-making. These are the specific domains where agency determines the final outcome.

While AI tools are able to generate increasingly fluent outputs, they currently lack human sense-making to determine meaning, context, or consequence. When the people using these tools struggle to make sense of the production or lack the confidence to question it, the output is often accepted uncritically or ignored entirely. These challenges stem from agency rather than technology.

When AI is introduced, people typically fall into one of two camps:

Embracing Experiments immediately Seeks feedback and iterates Shares what works with peers Asks questions even when uncertain

Waiting and Resisting Waits for formal training Cites busyness as a reason not to learn Avoids the tool quietly and carries on “business-as-usual” Accepts outputs without questioning them

These behaviours serve as a leading indicator of organisational health. The gravity of the "waiting" camp often outweighs the enthusiasm of the "embracing" few; if resistance remains the default, AI initiatives stay stalled at the pilot stage. Scaling only occurs when the culture shifts to make active experimentation the path of least resistance.

What This Means for Organisations

McKinsey’s 2025 Superagency in the Workplace report reveals that only 1% of organisations have reached true AI maturity, where systems are fully integrated and delivering meaningful results.

Reaching that top tier depends on the human ability to move past passive compliance and into active ownership. Leadership teams must evaluate whether their current culture provides the safety and motivation for people to join that 1%, or if the environment inadvertently rewards the "waiting and resisting" mindset that characterises the 99%.

At aAdvantage Consulting, our work in Human Capital and Culture Transformation starts from a simple premise that sustainable change is highly correlated to both organisation and people agency. We partner clients to transform this premise into performance, shifting teams from passive compliance to active ownership so that digital and AI transformation can finally scale beyond the pilot stage, into sustainable outcomes.

In the next article, we look at what personal agency actually looks like in practice; how it shows up in everyday decisions, and what distinguishes someone who takes ownership from another who waits to be told.