The Human Signals Leaders Ignore Before Governance Failures Occur
- Jul 26
- 9 min read
Written by Simer Dhillon, Executive Leadership Strategist
Simer Dhillon is the Founder and Chief Architect of SHARP™ Leadership Academy, a global platform redefining ethical performance systems for executives. She transforms leadership through measurable integrity, resilience, and presence.
Governance failures rarely begin with a broken policy. They often begin with human signals: silence, unclear accountability, normalized workarounds, suppressed concerns, and narratives that become stronger than evidence. In this article, Simer Dhillon explores how leaders can recognize these early warning signs and why human governance intelligence is becoming increasingly critical in the age of AI.
“Evidence should test a narrative. A narrative should never determine how evidence is created, interpreted, or accepted.” – Simer Dhillon
Why governance failure occurs
Organizations rarely discover governance failure at the moment it begins. By the time a serious issue reaches a board meeting, audit report, regulatory review, employee complaint, or public crisis, the conditions that allowed it to develop may have existed for months or even years. The warning signs were often there.
People stopped challenging decisions. Employees became cautious about raising concerns. Accountability became ambiguous. Difficult conversations moved into private channels. Informal workarounds became normalized. Leaders interpreted silence as agreement. Sometimes, a narrative becomes more powerful than the facts beneath it.
These are human signals. They may be among the most overlooked sources of governance intelligence inside an organization.
Silence is not the same as agreement
One of the most dangerous assumptions leaders can make is that silence means alignment. It does not. Researchers Elizabeth Morrison and Frances Milliken described organizational silence as a collective phenomenon in which employees withhold information about potential organizational problems. Their research explored how organizational conditions can create a shared perception that speaking up is unwise.
That distinction matters enormously for governance. A leadership team receiving little resistance may believe it has achieved alignment when what it has actually created is caution. A board hearing few concerns may believe controls are functioning when people closer to the problem have concluded that raising concerns will accomplish little or may carry consequences.
The absence of challenge is therefore not necessarily evidence of institutional strength. Sometimes, silence is data.
Governance is more than what is written
Organizations have become increasingly sophisticated at measuring what can be documented. They track compliance requirements, policies, controls, risk registers, performance indicators, cybersecurity incidents, financial exposure, and, increasingly, AI-related risks. All of these mechanisms matter.
But governance does not operate only through documents and controls. Governance operates through people making decisions under pressure. A company can have a strong code of conduct and still develop a culture where employees hesitate to challenge senior leaders.
It can have an escalation policy while employees remain uncertain about whether using it will damage their careers. It can publish organizational values while rewarding behaviour that contradicts them. It can establish responsible AI principles while teams deploy AI systems faster than governance processes can evaluate their implications.
This connects with a principle I have explored repeatedly in my work on leadership and governance: standards have little institutional value if they disappear when pressure arrives. The formal system can look healthy while the human system is communicating something entirely different. That gap deserves leadership attention.
Five human signals leaders should watch
1. Silence where challenge should exist
When consequential decisions consistently generate little disagreement, leaders should become curious. Do people genuinely agree? Or has challenging authority become too costly? Healthy governance requires constructive friction.
Organizations need people willing and able to say, “I see this differently,” particularly when decisions carry significant ethical, technological, financial, or reputational consequences. Creating a mechanism for speaking up is not enough. People must believe they can use it.
The G20/OECD Principles of Corporate Governance reinforce this distinction by emphasizing confidential whistleblowing mechanisms, board oversight of those mechanisms, and the ability to report unethical or unlawful behaviour without fear of retribution. Speaking up cannot simply be permitted in policy. It must be protected in practice.
2. Decisions without clear ownership
Listen carefully to the language surrounding difficult decisions. “We thought.” “They decided.” “Someone was handling it.” “We assumed it had been approved.” Ambiguous language can reveal ambiguous accountability.
This connects to another principle central to my work: clarity is not merely communication. It is infrastructure. If an organization cannot clearly identify who owns a decision, who has authority to challenge it, who must document it, and who is responsible for escalation, accountability can disappear precisely when it is needed most.
The OECD’s corporate governance principles similarly emphasize clear lines of responsibility and accountability throughout an organization. Good governance makes responsibility visible.
3. Informal workarounds becoming normal
Workarounds often begin innocently. A process seems too slow. A deadline is approaching. A client needs an answer. A new technology makes an old approval mechanism inconvenient. One exception becomes several.
Eventually, the workaround becomes the real operating system while the official process remains in the policy manual. This is where leaders must distinguish agility from governance erosion. If an organization repeatedly needs to circumvent its own processes to function, something requires attention.
Either the process needs redesigning or the behaviour needs correcting. Ignoring the contradiction creates institutional risk.
4. Bad news travelling slowly upward
One of the most revealing measures of institutional health is not how quickly good news reaches leadership. It is how quickly uncomfortable information does. If positive information travels rapidly while difficult information is filtered, softened, delayed, or discussed privately but never formally escalated, senior leaders may be operating with an incomplete version of organizational reality.
Governance becomes particularly vulnerable when the people closest to a problem have the least authority to change it. That is why safe escalation matters. The OECD recommends confidential mechanisms through which workers can raise concerns about unethical or illegal behaviour and protections against discriminatory or disciplinary consequences for doing so.
5. When narratives begin shaping the evidence
There is another human signal that deserves far greater attention: what happens when an institutional narrative becomes increasingly difficult to question and eventually begins influencing how evidence itself is created, gathered, or interpreted. Organizations need narratives. They help people understand strategy, identity, purpose, and change.
But narratives become dangerous when they begin replacing objective inquiry. A version of events may be repeated until repetition itself creates legitimacy. Contradictory information may receive less attention. People questioning the accepted interpretation may gradually be characterized as difficult, resistant, negative, or misaligned.
There is an even deeper governance risk. Evidence does not always emerge in a completely neutral environment. Once a conclusion or narrative has taken hold, circumstances may be selectively framed, interactions may be influenced, or people may be placed in situations where their predictable reactions are later interpreted as confirmation of what was already believed.
This becomes particularly concerning when people with less institutional power, including younger employees, early career professionals, interns, contractors, or employees who depend on the organization for future opportunities, are drawn into situations without understanding the wider context. They may act on partial information, follow direction, or respond naturally to circumstances that have already been framed for them. Their subsequent behaviour can then appear to provide independent support for a narrative they neither created nor fully understood.
This creates a critical governance distinction between evidence that independently establishes what happened and evidence produced within circumstances that may already have been influenced by a particular assumption, narrative, or desired conclusion.
Responsible leaders must therefore go beyond asking, "What evidence supports this narrative?" They must also ask:
How was the evidence produced?
Were the circumstances genuinely neutral, or could they have been influenced?
Did everyone involved have access to the same relevant information?
Were people knowingly or unknowingly being used to validate an existing assumption?
Were contradictory facts and alternative explanations given equal consideration?
Who controlled the process through which the evidence was gathered and interpreted?
Most importantly, did the conclusion emerge from the evidence, or was the evidence being interpreted through a conclusion that already existed?
These questions matter because institutional power can influence not only which narrative receives credibility, but also which information is collected, whose behaviour is observed, whose testimony is trusted, and which facts are ultimately considered relevant.
That does not mean every disputed narrative is deliberately manufactured. It means the integrity of the process used to establish institutional truth must itself be governable, transparent, and open to challenge. People, particularly those with less institutional authority, should never become instruments for manufacturing credibility around decisions or conclusions made elsewhere.
Evidence should test a narrative. A narrative should never determine how evidence is created, interpreted, or accepted.
AI makes human signals more important, not less
Artificial intelligence makes these questions even more urgent. AI is moving beyond generating content. AI systems and agents can increasingly recommend actions, prioritize information, influence decisions, trigger workflows, and perform tasks with varying degrees of autonomy.
This builds on an argument I have made previously about AI and leadership: the more powerful our technology becomes, the more consequential human judgment becomes. Technology does not eliminate accountability. It changes where accountability must be designed.
The U.S. National Institute of Standards and Technology’s AI Risk Management Framework organizes AI risk management around four functions: Govern, Map, Measure, and Manage. Importantly, NIST treats governance as a cross-cutting function and calls for clear organizational roles, accountability structures, ongoing monitoring, and defined human oversight.
That means organizations should not ask only: “Is the AI system working correctly?” They should also ask: “How are humans behaving around the system?”
Are employees questioning AI recommendations? Do they understand when human intervention is required? Can an early career employee challenge an AI-supported decision endorsed by a senior executive? Who owns the consequences when an automated workflow produces harm?
Does efficiency receive greater organizational reward than responsible judgment? When AI-generated information reinforces an existing institutional narrative, who verifies whether that narrative is actually true?
A technically sophisticated AI governance framework can still fail if the human environment surrounding it discourages challenge, obscures accountability, or rewards speed over judgment. Responsible AI therefore requires more than technical controls. It requires human governance infrastructure.
From compliance data to governance intelligence
This is why I believe organizations need to broaden how they understand governance maturity. Within the SHARP™ Governance Intelligence approach, I examine governance through three interconnected lenses:
70% – Operational evidence: What actually exists and operates: policies, decision structures, escalation mechanisms, accountability systems, controls, documentation, and evidence of implementation.
20% – Human signals: How people experience and behave within those structures: trust, challenge, willingness to escalate, leadership consistency, accountability behaviour, and responses under pressure.
10% – Standards alignment: Whether organizational practices remain connected to the ethical, regulatory, professional, and governance standards the institution claims to uphold.
These percentages do not mean human behaviour matters less. They reflect an important methodological principle: governance assessments should remain primarily grounded in verifiable operational evidence while still recognizing the human conditions that determine whether those systems actually work.
A policy can be verified. An organizational chart can be inspected. An escalation mechanism can be documented. But whether people trust the system enough to use those mechanisms requires another form of intelligence. Organizations need both.
The question leaders should ask earlier
Boards and executives frequently ask, “Are we compliant?” It remains an essential question. But another question may reveal problems much earlier, “What is our organization trying to tell us before the metrics do?”
Look at where employees hesitate. Look at where decision ownership becomes unclear. Look at which processes are routinely bypassed. Look at which risks everyone recognizes privately but few people raise formally.
Look at which narratives cannot comfortably be questioned. Look at whether those with the least institutional power can challenge those with the most. Look at whether leaders receive uncomfortable information early enough to act on it.
These signals should not automatically be treated as proof of misconduct. They are diagnostic intelligence. They tell leaders where to look deeper.
Governance failure has a prehistory
Major institutional failures can appear sudden from the outside. Internally, they rarely are. Before the investigation, there was often hesitation. Before the public controversy, there may have been an unresolved escalation.
Before accountability collapsed, responsibility may already have been unclear. Before a misleading narrative became accepted, someone may have known that the evidence told a different story. Before trust disappeared, people may have learned that speaking openly carried consequences.
That period before failure becomes visible is where leadership has its greatest opportunity. Organizations that learn to detect and interpret human signals can move governance from a retrospective exercise, explaining what went wrong, to a preventive capability that identifies where institutional integrity may be beginning to weaken.
The future of governance will not be built through more policies alone. It will depend on an organization’s ability to connect standards, operational evidence, human behaviour, accountability, ethical power, and AI-enabled decision-making into one coherent system.
Because sometimes the most important governance signal is not what an organization can measure. It is what its people have stopped saying.
If your organization is navigating AI adoption, accountability gaps, institutional trust, or governance complexity, now is the time to examine not only what your policies say, but what your people and systems are signalling.
Explore the SHARP™ Governance Intelligence approach and discover how stronger governance infrastructure can help identify risks before they become institutional failures.
Read more from Simer Dhillon
Simer Dhillon, Executive Leadership Strategist
Simer Dhillon is a leadership strategist and the Founder of SHARP™ Leadership Academy, a global platform integrating ethics, emotional intelligence, and performance systems for the modern workplace. Drawing on two decades in corporate finance and executive leadership, she developed the SHARP™ Framework (Standards, Honesty, Alignment, Resilience, Presence) to help leaders turn integrity into infrastructure. Her work blends business intelligence with emotional depth, empowering organizations to build cultures of measurable trust and sustainable success. Simer’s mission is to lead a new generation of ethically intelligent leaders who transform systems from within.
