The Illusion of Security: Why Your Internal Control Might Be a House of Cards

 

Once, during a consulting project, a manager told me something that sounded reassuring: "We have control for outgoing goods; it’s done every single day." However, as I dove into the process, I discovered the harsh reality behind those words. That "control" was merely an employee jotting down what left the warehouse in a notebook, no verification, no signature, and no oversight. In the manager's mind, the risk was covered; in practice, he had a hollow procedure, but not an effective control.

Many people confuse control with bureaucracy, and I often hear that "bureaucracy kills business." But that isn't true. 

Bureaucracy, in its essence, is not a bad thing. If we look at Max Weber’s theory, we understand that it was conceived as a form of human organization based on rationality, ensuring impersonality, a clear hierarchy, and meritocracy. It ensures that the process depends on the rule, rather than the mood of whoever is executing it. 

I always say that bureaucracy is simply the formalization of what is already working.

The real problem isn't bureaucracy itself, but "bureaucratic excess", when the ritual becomes more important than the result. 

An internal control system based on best practices, right sized to keep risk factors within acceptable levels, is essential for operational efficiency. It creates order and prevents the "rigidity" that often blinds an operation.

We know that internal control is an action designed to mitigate the cause of a risk before it materializes. It manifests in the "doing": in the review, the verification, the recalculation, and the careful approval. If there is no confrontation between "what should be" and "what is," the risk continues to walk freely through your company’s hallways.

For a control action to be more than just figurative, it requires four non-negotiable attributes:

• First, the Objective, which is the very reason for its existence and must target the risk factor. For example, if the risk is the use of incorrect labor hours, the objective is to ensure that the hours in the payroll system are consistent with the time-tracking system.

• Second, Practical Action, such as a data verification. It is vital to formalize how this action is performed so that a "prudent person" could re-perform it.

• Third, the Evidence, because a control without a trail is invisible. It could be a sign-off, a system log, or an email. To an auditor or an internal controls specialist, a lack of evidence implies the control does not exist. Finally, the Frequency, which must match the speed of the risk, whether it be daily, monthly, or per event.

 

When we talk about Control Modeling, the starting point is knowing the risk factor and the magnitude of what we are mitigating, allowing us to define the necessary attributes to keep the operation within the organization's risk appetite. 

These same attributes allow us to evaluate Design Efficiency, concluding whether the control has the theoretical capacity to mitigate the risk.

 Once we move to Effectiveness Evaluation—the actual control testing—the focus shifts to two crucial points: discipline, observing if the evidence exists within the defined frequency; and quality, re-performing the action to ensure it wasn't just a formality. 

This is why these attributes must be formalized, at the very least, in an Internal Control Matrix.

 It is also fundamental to understand that control is neither absolute nor infallible, as it is designed and executed by people. And where there are humans, there is vulnerability—whether through errors in judgment or, in grave cases, omission, fraud, and collusion. Therefore, the effectiveness of a control does not reside solely on paper, but in the discipline and ethical stance of its execution.

Recently, the market witnessed a testimony that serves as a stark warning. A former bank director admitted to signing documents without reading them. More than a momentary lapse, his confession revealed a void: despite heading the department, he did not perform monitoring or prevention duties.
Here we have the perfect "Window Dressing Control"—the evidence and frequency exist, but the execution of the action is null.

Unfortunately, this is not an isolated case; it is present in many corporations where management still mistakenly views control as mere bureaucracy.

To conclude, I leave you with a thought for reflection:

"The effectiveness of a control depends on action, evidence, and frequency, but it only stops being a house of cards when the integrity of the person executing it is greater than the convenience of simply signing."

Be happy!

Back to Basics: Why AI Won't Save an Audit Without a Solid Foundation

Happy 2026 Everyone!

As we kick off 2026, the buzz in the hallways of every organization is the same: Artificial Intelligence. We are talking about predictive analytics, automated reporting, and real-time monitoring. But as auditors, we need to have a moment of intellectual honesty.

There is no point in talking about Artificial Intelligence if we are still failing at the basics.

Disruptive innovation has indeed changed the way we work, but the pillars of a high-impact internal audit remain unchanged. Whether you are in the public or private sector, the value we provide isn’t just in the tools we use, but in the rigor of our methodology.

The "Shiny Object" Trap

It is easy to get distracted by the latest software, but a sophisticated algorithm applied to a flawed audit plan will only produce "high-tech" errors. High-impact auditing is born long before the first line of code is written; it is born from:

• Respect for Standards: Adhering to professional norms is what gives our work legal and institutional weight.

• Rigorous Planning: Understanding the object of evaluation, defining clear scopes, and setting realistic timelines.

• Quality of Evidence: Collecting evidence that is not just abundant, but sufficient, reliable, relevant, and useful.

Returning to the Foundations

For a truly effective audit, it is essential to "work the basics." This means mastering the core cycle of our profession:

1. Planning with Purpose: Truly knowing the entity and its risks before defining objectives.

2. Execution with Precision: Applying the right techniques—be it physical inspection, circularization, or analytical review—and documenting them in flawless working papers.

3. Communication with Impact: Reporting not just "what happened," but providing a professional opinion that helps management improve through actionable recommendations and consistent follow-up.

Looking Ahead to 2026

My invitation to you this year is to review your foundations. Masterfully executing the basics is, ironically, the greatest competitive advantage an auditor can have in a world obsessed with automation.

AI can process data, but it cannot replace the professional judgment of an auditor who knows how to apply a materiality lens or evaluate the "design and discipline" of an internal control system.

Let’s embrace the future, but let’s keep our feet firmly planted on the bedrock of our profession.

Be happy! 

The Critical Role of Artificial Intelligence Governance in a World Undergoing Profound Transformation

 

I have been participating in several working groups where we discuss the impact of artificial intelligence (AI) on organizations and society. And I must confess: the deeper I dive into this subject, the more I realize we are living through a moment of transformation far deeper than it seems at first glance.

In recent years, technology has evolved at an unprecedented speed. Innovations that once seemed distant are now knocking on our doors, radically changing how companies operate and compete. If we once spoke of isolated disruptions, we now live in a state of continuous disruption, spanning all areas and directly challenging corporate strategies.

And who is at the center of this revolution? Artificial intelligence, of course. But not AI alone. Technologies like autonomous agents, quantum computing, and neuromorphic computing are reshaping the fabric of organizations and, consequently, of society itself.

What strikes me most is that, despite all this, many professionals and companies have yet to grasp the depth and speed of these changes. And that’s concerning. It could compromise the sustainability of organizations and even the relevance of many professionals.

This new scenario demands more than enthusiasm for innovation. It demands responsibility, structure, and governance. And this is where AI governance comes in—as a viable, necessary, and urgent path to ensure that all this transformation is sustainable, ethical, and strategic.

We are entering an ecosystem of emerging technologies that, when combined, have the potential to completely reshape the fabric of society. More than just new tools, they represent new paradigms.

Much has been said about generative artificial intelligence, which is undoubtedly a milestone. But it’s only the doorway. We are witnessing the convergence of several emerging technologies that together have the power to radically change how we live, work, and relate to one another.

I want to share some of this reflection with you, starting with understanding what is truly happening around us.

Autonomous Virtual Agents

Autonomous virtual agents are not just smarter chatbots. They are systems capable of understanding objectives, making decisions, and executing tasks without continuous human supervision.

These agents are already being tested in financial negotiations, medical diagnostics, and even customer service learning, interacting, and adapting in real time.

Soon we will have our own virtual agent capable of performing simple tasks like receiving our emails, reading, assessing, prioritizing, deciding, and responding as if it were us even carrying out complex operations, such as defining products based on actuarial calculations.

Their ability to operate in complex and dynamic environments raises crucial questions about control, responsibility, and social impact.

Autonomous Robotic Systems

Industrial automation has adopted a new meaning with autonomous robotic systems, where robots not only follow instructions, but they also learn from their environments, correct their paths, collaborate with each other, and make decisions based on data.

Tesla’s Optimus is an example of this, expected to hit the market by 2026, at the price of a vehicle. In sectors such as logistics, healthcare, agriculture, defense, and space, these systems are replacing human labor in critical tasks raising significant questions about employment, ethics, and safety.

Quantum Computing

The promise of quantum computing is simple yet monumental: solving problems that would take a traditional supercomputer millions of years—in just minutes or seconds.

This could transform areas like climate modeling, molecular simulations for new drugs, logistical optimization, and especially cybersecurity.

IBM has already launched the Q System, a commercial quantum computer. Google has achieved “quantum supremacy” by performing a calculation on a quantum computer that would be impossible for classical supercomputers in a reasonable time. Microsoft is advancing in qubit technology, having developed a new quantum chip capable of solving large-scale complex problems.

With this power come significant risks, such as the potential to break encryption systems that underpin the modern internet, exposing sensitive data from governments, companies, and citizens.

Neuromorphic Computing

Although not new—Misha Mahowald and Carver Mead developed the first silicon retina and cochlea, as well as the first silicon neurons and synapses in the 80s neuromorphic computing gain renewed relevance with generative AI.

Inspired by the functioning of the human brain, neuromorphic computing seeks to create systems with learning and adaptation capabilities that closely mirror biological cognition.

This represents a major leap toward the creation of truly autonomous AI, capable of reasoning with context, memory, and emotion.

But it also represents a turning point: how do we regulate machines that think similarly to us?

Observe that what makes this moment unique is not the emergence of one disruptive technology, but the convergence of several.

When autonomous agents operate using neuromorphic neural networks, supported by decisions optimized through quantum algorithms, within robotic ecosystems, we are undoubtedly facing a new form of systemic intelligence—one that, if not properly governed, could surpass our control, with unpredictable consequences.

In 2024, at an AI event, a presenter stated that soon we would have three types of agents operating in companies: human agents, hybrid agents, and android agents.

At the time, I thought, “This speaker is watching too many sci-fi movies.” But today I see how mistaken my view of technological evolution was. Not as an excuse but understand that I’m almost a "time traveler” when I began working half a century ago, the most advanced technology was a manual typewriter, or a communication device called the Telex.

But let’s set aside the nostalgia and continue our reflection. 

Thus, it's clear that beyond the fascination with new technological possibilities, there’s a reality that organizations cannot avoid: the way they operate, protect themselves, and provide accountability is being profoundly reconfigured—so rapidly that it’s difficult to process, adapt to, and integrate innovations into daily operations.

And since companies are made up of people, all this deeply and continuously affects individuals’ lives, requiring them to break paradigms constantly. This contributes significantly to professional burnout and rising depression levels.

Corporate structures are being redefined, and this doesn’t only affect IT it impacts the entire operational ecosystem. Every department is being affected, without exception.

Compliance, risk management, and internal audit—traditionally pillars of corporate governance are directly impacted by this new disruptive ecosystem. Let’s examine:

Compliance

With the rise of autonomous agents and real-time decision-making systems, ensuring legal and ethical compliance is no longer a matter of simply “checking processes.” It now demands continuous monitoring, a deep understanding of the technologies involved, and the ability to respond to unforeseen events.

What happens, for instance, when an autonomous AI makes a biased or unethical decision? How can we ensure systems comply with regulations that are still being formulated?

Risk Management

In today’s landscape of exponential innovation, corporate risk management faces one of its greatest challenges: anticipating the unpredictable.

Technologies like autonomous agents, quantum computing, and neuromorphic systems introduce variables that didn’t exist a few years ago—and often aren’t even recognized as risks until they’ve already materialized.

The traditional risk management model—based on static cycles of identification, analysis, response, and monitoring—was already showing signs of exhaustion and now must be completely reimagined. It lacks the agility and adaptability to handle emerging risks that evolve in a matter of weeks, days, or even hours.

Risk is no longer a possibility—it’s a certainty at some point in the journey. The real differentiator now is the speed and intelligence of the response. This demands new organizational capabilities.

Internal Audit

Internal audit, long the guardian of compliance and efficiency, must now also serve as an interpreter of technological complexity.

With increasingly automated processes and decisions made by autonomous systems, auditing the "who did what" requires understanding algorithms, data flows, and machine learning logic.

More than identifying failures, auditing now requires anticipating risks, evaluating efficiency considering new innovations, assessing ethical impacts, and verifying whether digital governance principles are being upheld.

So, the central question is: How will we, and our organizations deal with all this?

In my view, there is no single answer. But one thing is certain: the first step involves a structured approach to effectively manage this technological disruption in a sustainable way, which we can call artificial intelligence governance.

AI governance is not just a control strategy, it is a foundational approach to ensure digital transformation occurs in alignment with corporate, societal, and ethical interests.

In times of rapid and unpredictable innovation, it serves as the backbone for managing disruption sustainably creating a framework that guides organizations not just to innovate, but to innovate with responsibility and long-term vision.

For our reflection, I believe AI governance must address the following key areas:

Defining Responsibilities

AI governance sets clear responsibilities within the organization. Who is accountable for the ethical and safe use of technology? How do we ensure automated or AI-assisted decisions follow company guidelines and legal standards?

Creating an AI governance committee, for instance with executives from IT, compliance, legal, and ethics, ensure decisions are made in a coordinated and informed way, without overwhelming any single department.

Additionally, governance determines how responsibilities align with strategic objectives. Every new AI project should be evaluated not only for its innovative potential, but also for its strategic, ethical, and regulatory impact.

Committing to AI governance means that, while the organization explores new technological frontiers, it also maintains control over the consequences of innovation.

Defining Security Standards

As technologies advance, security becomes a critical issue—not just in terms of data protection, but also regarding the integrity of automated decisions and system reliability.

AI governance establishes the necessary security standards to protect both sensitive data and autonomous systems. This involves implementing advanced cybersecurity mechanisms and protocols to ensure AI makes decisions that are secure and aligned with the organization’s values.

Preventing bias, ensuring algorithm transparency, and auditing automated decisions are all essential governance practices to ensure technologies are not only effective but also safe and fair.

Operational Formats and Monitoring

Governance also defines the operational structure of AI within the organization—creating frameworks for the development, integration, and management of intelligent systems.

AI implementation must be transparent, continuously monitored, and adjusted as technology evolves.

AI monitoring systems are essential to ensure that even when systems make autonomous decisions they remain within established boundaries.

In addition, AI governance demands ongoing monitoring to detect failures, errors, or unwanted behavioral changes, ensuring that unexpected risks don’t arise.

This monitoring must be integrated into strategic corporate management, aligning technological innovation with organizational goals and culture, so that AI contributes effectively to sustainable and ethical growth.

Note that this topic is much broader and far from exhausted here. Organizations through their Boards of Directors and/or Executive Management must address it quickly, seriously, and assertively.

They must take a leading role in structuring robust governance that permeates the entire organization and this is now urgent and non-negotiable.

I leave you with a thought provoking question for reflection:

How are you and your company addressing this issue?

The journey is just beginning. And despite all the challenges, what truly matters in the end is that we continue to find meaning, build together, and be happy
with ethics, awareness, and purpose.

Comments are welcome! Be happy.

This article was written with the help of human intelligence!