Setting the Stage
That quote from Antonio Sánchez Gómez captures a reality many of us in data have lived: organizations where governance “exists”, but only in slide decks, policy documents and the heads of a few experts. It is governance you can recite, but not governance you can feel.
My nuance was this: governance should not live in institutional memory. It should live in institutional muscle memory.
As long as governance has to be remembered, it is fragile. When it is embedded in processes, metadata structures, approval workflows and system constraints, it becomes behavioral rather than declarative.
In other words, it stops being something people comply with, and starts being something the organization does almost without thinking.
From Memory to Muscle Memory
In neuroscience, muscle memory is not really in the muscles. It is in the neural circuits that automate skills through repetition: typing without looking at the keyboard, playing an instrument or keeping correct form in an exercise without reciting the instructions in your head.
Two aspects of physical muscle memory are highly relevant for governance:
A skill becomes automatic when practice is consistent and guided by feedback, including successes, errors and corrections.
Once consolidated, it can be regained surprisingly fast after a period of detraining.
In an organization, the muscles are not biceps. They are workflows, metadata models, tools and the decisions people make repeatedly.
Organizational muscle memory appears when:
The same types of decisions are taken in the same way across teams.
Exception paths are designed, documented and used consistently.
Systems make the right path structurally easier than the workaround.
That does not happen organically. It has to be deliberately designed.
Gestalt Moments and the “Now I See It” Effect
When I tried to explain this to a leadership team that wanted governance results “yesterday” and to data users pushing back on having to memorize processes that would soon change, I realized this was not a slideware problem. It was an insight problem.
Gestalt psychology describes insight as that “aha” moment when we mentally restructure a problem and suddenly see a solution or pattern with striking clarity. It is not linear or gradual. Often it comes after a period of impasse, frustration and feeling stuck, followed by a sudden reorganization where understanding feels immediate and emotionally clear.
In data governance, this insight shows up when:
A data user stops seeing governance as bureaucracy and starts seeing it as a way to avoid rework, firefighting and endless “where did this number come from” threads.
Management stops seeing governance as friction and starts seeing it as the only way to scale data driven decisions without multiplying risk.
It is not about memorizing a process. It is about internalizing a mental model: this is how we treat data here.
That shift, from control for compliance to frictionless safety and clarity, is deeply Gestalt. The underlying reality is the same, but the way we perceive and organize it changes completely.
Why Governance Gets Stuck in Institutional Memory
The adoption of data governance programs has risen significantly. Recent industry insights suggest that around seven in ten organizations now report having a formal data governance program, up from roughly six in ten only a couple of years ago. But that does not mean governance is truly operating in the day to day.
Several surveys and reports highlight recurring patterns:
Governance is implemented primarily to satisfy regulation, not to accelerate decisions or improve quality.
Around two thirds of organizations name governance as one of the top challenges in making artificial intelligence work at scale, despite having formal programs in place.
Implementation remains stuck in policies, committees and one off trainings, with limited translation into process and tooling changes.
The result looks like what Oksana Chekanova calls the silent drift of risk boundaries. Boards define risk appetite, it becomes a policy, training is done, and then day to day exceptions and shortcuts gradually move the limit, without anyone joining the dots explicitly.
The memory of the initial decision dissolves. Governance becomes a narrative, not a reflex.
That is the trap of institutional memory. It depends on who was in the room, who read the document and who remembers how we decided this. It does not scale. It does not survive rotations, reorganizations or crises.
Designing Governance as Muscle Memory
Back to Antonio Sánchez Gómez ’s question.
In your experience, what tends to embed muscle memory most effectively: tooling constraints, workflow design or accountability pressure?
My short answer. You need all three, but the order matters.
Tooling. Make the Right Path the Easy Path
Tooling constraints, when designed well, are like good technique at the gym. They protect you even when you are tired.
Modern data catalogs and governance platforms that expose ownership, definitions, quality and permitted use clearly reduce the cognitive load of remembering rules.
Templates and data models that include mandatory governance fields such as owner, classification, purpose and retention turn compliance into a natural step in creation, not an extra checklist.
Access controls, masking policies and personal data rules enforced at the data or semantic layer prevent every individual team from reinventing controls from scratch.
Tools do not replace culture. But they prevent culture from breaking at the first tight deadline.
Workflow Design. Teaching by Doing
Workflow is where theory becomes muscle memory.
Approval flows with clear decision rights, specifying who decides what, in which timeframe and based on which minimal information, reduce the need for “ask the one person who always knows”.
Embedding governance steps into existing processes, such as source onboarding, product launches and artificial intelligence model changes, avoids the parallel process syndrome that nobody follows.
Training on real cases, not artificial examples, helps users experience the insight. They see how aligned definitions or better metadata avoid errors and reconciliation drama.
This is where I have seen the most impact. Identify data users who have already internalized governance and pair them with colleagues who are still on the journey. It is vicarious learning. People copy not just the steps, but the way of thinking.
Accountability. Pressure That Clarifies, Not Paralyzes
Accountability pressure by itself creates fear, not memory. Well designed, it creates urgency and coherence.
Clearly defined roles such as owners, stewards and custodians tied to real objectives, not just titles, align incentives for data caretaking.
Measuring governance by initiatives enabled, rework reduced or quality improved, not only by number of policies or controls, reframes success.
Making trade offs visible, including why an exception is granted, under which conditions and with what learning, feeds a collective memory of decisions, not just a library of rules.
The key is that accountability is not only about who signs. It is about how we decide here when something does not fit the manual.
Training Data Fluency, Not Governance Fear
In my experience, the real inflection point comes when we stop asking people to memorize governance and instead help them internalize it.
Several practices have proved effective.
Continuous training programs, not a single governance day, that connect governance concepts to real business problems, supported by live tools and real datasets.
Pairing or governance buddying between data users with high data fluency and colleagues who are still building it, so governance is transmitted through observation, shared work and daily conversation.
Systematic use of metrics such as data quality scores, time spent cleaning data and access friction as feedback loops. When cleaning time goes down and trust goes up, people start to see the link between governance and delivery speed.
This directly impacts organizational data fluency. More people can read, question and create with data safely because context is in the system, not in a forgotten document or the memory of a single steward.
On the ground, it looks simple. Fewer tickets of “I do not trust this number”. Fewer sterile debates about definitions. Fewer reconciliation meetings. And more energy invested in designing new questions, products and artificial intelligence use cases.
From Remembered Governance to Embodied Governance
Most organizations already have governance. What they lack is practice.
Institutional memory is fragile, personal and non transferable. Organizational muscle memory is the opposite. It is structural, repeatable and resilient to people changes and context shifts.
To get there, well written policies are necessary but not sufficient. You need:
Tools that make the right path the easiest one to follow.
Workflows that teach, repeat and reinforce the desired decision patterns.
Accountability that turns decisions and exceptions into shared learning, not isolated episodes.
And above all, you need to recognize that real adoption depends on that Gestalt style insight. The moment when someone stops seeing governance as an external imposition and starts seeing it as the natural way of working with data.
That is the day governance stops being memory.
And starts being muscle.