The organization has trained everyone in how to speak up.
There is a workbook, a short video, a discussion guide, and a line in the leadership principles about candor. Yet in the meeting where speaking up matters, no one does it.
The silence is not explained by a lack of knowledge.
The junior employee knows the language. The manager has completed the training. The policy is visible. But the meeting is already late, the senior leader has announced a preference, the agenda allows no time for dissent, the room is arranged around a single presenter, and everyone remembers what happened to the person who challenged the last launch.
If we audit only the people, we will diagnose the wrong system.
A practice architecture audit examines the arrangements that make a practice easier, harder, normal, risky, thinkable, or almost impossible in a particular site. The method is grounded in the theory of practice architectures developed by Stephen Kemmis and colleagues. That theory analyzes practices in relation to cultural-discursive, material-economic, and social-political arrangements that enable and constrain sayings, doings, and relatings.
The exact phrase "practice architecture audit" does not name one standardized validated instrument. In this library it describes an applied analytic method based on an established practice-theory framework.
That distinction protects the page from pretending that a checklist is science merely because the theory beneath it is serious.
Practices live in sites
Practice architectures theory begins from a deceptively simple idea: practices are not free-floating behaviors.
What people say is made possible and constrained by language, discourses, categories, and shared meanings. What people do is shaped by material conditions such as time, space, tools, money, schedules, and physical arrangements. How people relate is shaped by power, solidarity, authority, roles, and social structures.
Kemmis and colleagues describe these as cultural-discursive, material-economic, and social-political arrangements.
A practice therefore becomes intelligible only in its site.
Take reflective decision-making. An organization may value it explicitly. Cultural-discursive arrangements affect whether uncertainty has legitimate language. Material-economic arrangements determine whether there is time to review. Social-political arrangements shape who is allowed to challenge whom.
If any one of these architectures works against the practice, training the individual harder may produce little change.
What an audit asks
A practice architecture audit should begin with one practice, not an entire organization.
"Culture" is too large. "Better collaboration" is too vague. A useful target is something observable and consequential: raising risk before deployment, giving feedback after error, reviewing AI-generated analysis, conducting learning debriefs, protecting focused work, or involving affected users in a design decision.
Then the audit asks three linked sets of questions.
Cultural-discursive: What words, concepts, stories, assumptions, metrics, and professional languages make this practice sayable? Which meanings make it difficult to name?
Material-economic: What time, space, tools, staffing, workflow, money, data access, and technical infrastructure make the practice possible? What material arrangements make it impractical?
Social-political: What roles, authority relations, dependencies, incentives, solidarities, and sanctions shape who can act, who can refuse, and whose interpretation counts?
The power of the method lies in refusing to reduce these dimensions to attitude.
People can believe in a practice and still inhabit architectures that block it.
Evidence collection should follow the practice
An audit becomes weak when it consists only of interviews about what people think happens.
Self-report matters, but practices leave multiple traces.
A robust audit can combine observation, meeting artifacts, schedules, policies, physical or digital workflow, decision logs, communication patterns, role definitions, budget allocation, tool permissions, and interviews across positions. The goal is not surveillance. It is triangulation.
If leaders say reflection is protected, calendars can show whether review time exists. If a team says dissent is welcome, decision records can show whether objections change outcomes. If an AI policy requires human review, workflow observation can reveal whether the reviewer has time, expertise, and authority to do anything other than approve.
Evidence should remain close to the practice being analyzed.
This also supports credibility checks. Different participants occupy different positions in the architecture. A manager may experience autonomy where a contractor experiences dependence. A policy author may see a permission that a frontline worker knows is unsafe to use.
One account should not stand in for the site.
From audit to change hypothesis
The method should not jump directly from observation to solution.
An audit produces hypotheses about which arrangements are enabling or constraining a practice.
Suppose a team rarely challenges AI-generated recommendations. The first explanation might be overtrust in automation. The audit may find a different architecture: the recommendation is presented only after senior leadership has already chosen a direction; reviewers receive it minutes before the meeting; the interface hides source uncertainty; and performance metrics reward rapid implementation.
The problem is not merely individual automation bias.
The intervention might involve changing the timing of review, making provenance visible, altering decision authority, or separating recommendation generation from final framing. Training may still help, but it is no longer asked to carry a structural problem alone.
This is where practice architecture connects to Inner Technology. Human capacities develop and operate inside arrangements. If an institution wants discernment, attention, relational maturity, or ethical judgment, it has to examine whether its site architectures allow those capacities to become practices.
Swedish and Nordic relevance
Practice architecture research has direct Scandinavian relevance rather than requiring a decorative Nordic citation.
A published study of middle leadership in early childhood education examined practice architectures in a Swedish city district. The researchers used the framework to analyze how cultural-discursive, material-economic, and social-political conditions shaped middle-leading practices across preschool settings.
The importance of the example is methodological. The framework can make local institutional conditions visible without assuming the same arrangements operate everywhere.
Nordic workplaces and public institutions often carry strong formal commitments to participation, learning, equality, and trust. An audit can examine the difference between those formal commitments and the architectures through which daily practice actually occurs.
It should not assume that national culture determines the result.
Scoring is optional and risky
An audit does not need a maturity score to be useful.
Qualitative mapping may be more faithful when the goal is diagnosis. Teams can identify enabling and constraining arrangements, contradictions between dimensions, and leverage points for change.
If scores are introduced, the measurement model must be explicit. What does a "4" on social-political enablement mean? Can two raters apply the anchor consistently? Are dimensions independent? Does a higher score predict better enactment of the target practice? How are differences in perspective handled?
Without those answers, numbers create false precision.
The safest Institute default is descriptive evidence plus clearly bounded rating anchors only when a decision genuinely requires comparison over time or across sites.
The audit is primarily an inquiry method.
Relationship to practice architecture
The library also contains Practice Architecture as a conceptual entry. The distinction is important.
Practice architecture describes how developmental practices can be intentionally structured and supported. Practice architectures, in the Kemmis tradition, are the site-based arrangements that prefigure practices.
The terms overlap linguistically but should not be treated as identical.
A Practice Architecture Audit uses the established practice-architectures lens to examine conditions around a defined practice. An Institute practice architecture may then be redesigned in response. The audit diagnoses the site; the designed architecture organizes an intervention.
Keeping these layers separate prevents conceptual circularity.
Audit quality depends on the unit of practice
The most common failure in this kind of analysis is choosing a target that is too large to observe.
"Leadership culture," "innovation," "trust," and "wellbeing" are important topics, but they are not single practices. An audit becomes sharper when it names an activity with participants, purpose, sequence, site, and consequence. "How product teams escalate evidence of user harm before launch" is auditable. "Ethical culture" is a field requiring many such observations.
This specificity changes sampling. Researchers can observe several enactments of the practice across teams, shifts, or sites. They can compare what people say should happen with what actually happens. They can identify architectures that recur and those that are local exceptions.
It also makes change evaluation possible. If an intervention alters meeting timing, decision rights, source visibility, or escalation channels, the audit can examine whether the target practice changes rather than relying on a broad culture survey.
Reflexivity matters because the auditor also enters the site. Interview questions introduce categories. Observation can change behavior. Senior sponsorship can make participants cautious. An Institute protocol should therefore record who conducted the audit, what access they had, how participants were recruited, what evidence was unavailable, and where interpretation depended heavily on one perspective.
Qualitative rigor is not achieved by pretending the observer is absent. It is achieved by making the observer's position inspectable.
From architecture to leverage
An audit is most useful when it identifies a small number of changeable arrangements rather than producing a catalog of everything wrong with the organization.
Some constraints are foundational. If reviewers lack authority to stop a deployment, more reflective language will not create meaningful oversight. If there is no time allocated for debriefing, a learning norm will remain aspirational. If a metric punishes the behavior a value statement requests, the metric is part of the architecture.
The method therefore ends with leverage hypotheses, each tied to evidence and an expected change in practice. Those hypotheses can then be tested rather than treated as recommendations made true by the audit itself.
Contradictions are often the most useful finding
The three kinds of arrangements do not always point in the same direction.
An organization may have language that strongly supports autonomy while its material workflow requires approval for every meaningful action. It may fund generous learning time while social hierarchy makes admitting uncertainty dangerous. It may create egalitarian meeting rituals while performance incentives reward individual competition.
These contradictions should not be averaged away. They often explain why change programs feel sincere and ineffective at the same time.
An audit can therefore code not only enablers and constraints but tensions among arrangements. A practice becomes especially unstable when one architecture invites it and another punishes it. That pattern is often more actionable than a global score because it identifies the point where institutional messages become behaviorally incoherent.
An audit should follow consequences, not only arrangements
Practice architectures can make a pattern intelligible without proving what that pattern causes. An audit therefore becomes stronger when it follows the practice forward. If meeting structures change, does speaking distribution change? If protected review time is introduced, are consequential decisions actually challenged more often? If a new material tool removes friction, what capacities are strengthened and what forms of attention disappear with it?
This creates a second pass after description: consequence tracing. The team identifies a small number of changes in sayings, doings, or relatings that should become visible if the proposed architectural intervention matters, then returns to observe. The purpose is not to turn practice theory into a simplistic experiment or pretend every workplace variable can be isolated. It is to prevent the audit from ending with an elegant diagram. A useful audit creates a corrigible hypothesis about the site and then lets the site answer back. In that sense, the method is less like certification than disciplined institutional perception: notice the conditions, alter them carefully, and remain available to what actually changes.
What the evidence does not show
The theory of practice architectures does not validate one standardized Practice Architecture Audit instrument.
It does not show that every problem can be explained by three categories, or that cultural-discursive, material-economic, and social-political arrangements are statistically independent dimensions. They are analytic distinctions within a practice theory.
Case studies using the framework can show how it supports interpretation in particular sites. They do not establish that an Institute audit will produce reliable scores across organizations unless that scoring system is separately developed and tested.
The framework also does not remove individual agency. People interpret, improvise, resist, reproduce, and transform practices. Nor does attention to local arrangements eliminate wider forces such as law, market structures, political economy, historical inequality, or technological infrastructure.
Finally, a practice architecture audit should not be used to turn every human difficulty into an organizational-design problem. Some constraints are personal, clinical, developmental, material, or outside the institution's control. The point is not to replace one reductionism with another.
The room teaches the practice
Return to the silent meeting.
The training was real. The value statement was sincere. The people may even believe in candor.
But the room teaches too.
The clock teaches. The hierarchy teaches. The agenda teaches. The memory of punishment teaches. The interface teaches. The distribution of authority teaches.
A practice architecture audit makes those teachers visible.
Its purpose is not to excuse people from responsibility. It is to place responsibility where it can actually act: in the person, the relation, the material arrangement, the language, the rule, the design, and the institutional conditions that together decide which capacities become possible in practice.
Research references
Kemmis, S., Wilkinson, J., Edwards-Groves, C., Hardy, I., Grootenboer, P., & Bristol, L. (2014). Changing Practices, Changing Education. Springer. https://doi.org/10.1007/978-981-4560-47-4
Mahon, K., Francisco, S., & Kemmis, S. (Eds.). (2017). Exploring Education and Professional Practice: Through the Lens of Practice Architectures. Springer. https://doi.org/10.1007/978-981-10-2219-7
Kemmis, S. (2019). A Practice Sensibility: An Invitation to the Theory of Practice Architectures. Springer. https://doi.org/10.1007/978-981-32-9539-1
Grootenboer, P., & Edwards-Groves, C. (2024). The Theory of Practice Architectures: Researching Practices. Springer. https://doi.org/10.1007/978-981-99-7350-7
Ronnerman, K., Grootenboer, P., & Edwards-Groves, C. (2017). The practice architectures of middle leading in early childhood education. International Journal of Child Care and Education Policy, 11, Article 8. https://doi.org/10.1186/s40723-017-0032-z