To understand where we are going, we must first look at where we have been.

Loops are everywhere it seems. They appear in idiomatic speech, as in ‘close the loop‘, ‘keep me in the loop‘, ‘I’m out of the loop‘, ‘let’s loop back around on this‘, ‘they were thrown for a loop‘. (Jargon bingo cards may include one or more of these). Loops also play a significant role in nonprofit organisations: mostly beneficial, but sometimes also harmful.

Good Loops

Several kinds of beneficial loops can be identified, and some of these are outlined below.

In systems thinking – pioneered by Peter Senge and grounded in the foundational work of Donella Meadows in Thinking in Systems – reinforcing (positive) loops amplify change, while balancing (negative) loops seek stability or a specific goal by resisting change. These are often used in causal loop diagrams that map out the relationships between variables in a system.

In learning and decision-making, various loop models describe how individuals or organisations process information and adjust their behaviour.

• Double-loop Learning (Argyris) moves beyond fixing the problem (single-loop) to questioning the underlying assumptions or goals that created the problem (double-loop). A later iteration augmented this model with Triple-loop Learning, as illustrated in the chart below)
• The continuous decision cycle characterised as an OODA Loop (Observe, Orient, Decide, Act) by John Boyd promotes rapid response to changing environments.
• The PDCA loop (Plan, Do, Check, Act) devised by W. Edwards Deming offers an iterative approach to continuous improvement and problem solving.

These models are examples of ‘reflective practice loops‘ – structured, cyclical processes designed to turn experience into learning through continuous analysis. Other key models include:

• Gibbs’ 6-stage cycle – description, feelings, evaluation, analysis, conclusion, action plan
• Kolb’s 4-stage experiential learning cycle – Concrete Experience (doing/feeling), Reflective Observation (watching/reviewing), Abstract Conceptualisation (thinking/learning), and Active Experimentation (planning/trying)

These tools aid in deep learning by reviewing actions, questioning assumptions, and planning improvements. Such models may be used for self-development, or to help frame mentoring discussions, amongst other purposes.

In the context of project management and Agile methodologies, ‘closing the loop’ serves as a vital metaphor for the successful completion of a phase, sprint, or project. This cycle is effectively ‘opened’ at the moment goals or objectives are initially set, creating a focused period of activity aimed at a specific outcome. The loop is only considered ‘closed’ once those objectives have been achieved, resolving the ‘open loop’ and preventing the mental energy drain that characterises unresolved tasks. This project-based cycle functions as a practical application of the PDCA (Plan, Do, Check, Act) loop, where the final stages verify that the initial plan was fulfilled before the cycle is completed. Such disciplined closure ensures that a team is engaging in intentional iteration rather than simply repeating tasks without a clear ‘base case’ or exit condition.

‘Open loops‘ are also used by various disciplines to describe unresolved issues or ‘models of reality’ that drain mental energy until they are closed or resolved (closed loops).

A recursive loop is a programming technique where a function calls itself, either directly or indirectly, to repeat a task until a specific base condition is met. It acts as an alternative to iterative loops (for/while) and is useful for traversing complex, nested structures like trees or files. It requires a base case to prevent infinite loops and stack overflow errors. These loops aim to optimise code efficiency by reducing redundancy for repetitive tasks. (Note: NASA and Microsoft are examples of organisations that ban recursive code due to risks of stack overflows, high memory usage, and other factors).

This blog uses a temporal loop as its header image to describe the relationship between past, future, and present perspectives that inform all of the choices we make.

‘Loopy’ thinking

Some loops have negative connotations.

Just as computer code requires a ‘base case’ to prevent a crash, our mental patterns require an “exit condition” to avoid emotional exhaustion. When we are ‘stuck in a loop’, there is a sense in which we are continuing to apply the same (old) approach to an ever-changing situation, with less and less effectiveness.

In psychology, looping thought patterns are repetitive, involuntary mental cycles where the mind fixates on a specific worry, question or memory. Such self-critical, anxious or fearful thoughts can lead to emotional exhaustion and elevated stress levels.

Another psychological example concerns conflict loops. These are recurring, predictable, and unproductive argument patterns where two parties (usually couples) rephrase old hurts, resulting in emotional detachment and unresolved tension. These loops occur when surface issues (like chores) mask deeper, underlying unmet needs (like feeling ignored), leading to defensive, circular fighting. These patterns usually remain stuck because both parties feel they need to win the argument, and this prevents them for addressing the root of the problem. Similar patterns can be discerned between opposing ‘sides’ within teams, between organisations, and between nations.

One example of ‘loopy thinking’ is the use of ‘circular reasoning‘, where a thought pattern or argument structure includes the conclusion in its premise. This logical fallacy, also called ‘begging the question’, often prevents progress, causes overthinking, and reinforces irrational beliefs or anxiety. In a nonprofit context, an example might be “You must comply with the policy because our policy requires it”.

The Governance Loop: AI and Human Accountability

In the modern nonprofit landscape, the integration of AI introduces a complex new cycle: the human-technology feedback loop. While AI offers a significant opportunity to widen analytical support and generation of options, it also presents a “seductive” temptation to prioritise speed over depth. The “deeper risk” for organisations is not just technical error, but a drift toward outsourced judgment where the responsibility for decisions is blurred.

To counter this, nonprofit leaders must practice disciplined use, treating AI as an aid to reasoning rather than a displacement of human interpretation. In governance terms, this means directors and officers must remain both “in the loop” — actively reviewing AI-assisted work products – and “over the loop” – supervising the broader reasoning environment and governance structures. Ultimately, AI-assisted checking is merely a prompt for human review. Accountability for “moral seriousness” and “responsible judgment” remains an inherently human burden that cannot be transferred to a machine.

Mastering the Loop

Recognising that loops are omnipresent is only the first step; the real challenge lies in orchestrating them. To make effective use of ‘good loops’, we must move from passive repetition to intentional iteration. This involves adopting structured models like the PDCA loop for continuous improvement or the OODA loop for agility. Furthermore, true growth occurs when we transition from single-loop problem solving to double-loop learning, where we stop merely fixing symptoms and begin questioning the underlying goals – a practice essential when governing AI to ensure it serves ‘institutional trust’ rather than just ‘compressed reflection’.

Conversely, remedying ‘bad loops’ requires a high degree of pattern recognition and the courage to reclaim responsibility. Whether we are closing ‘open loops’ to preserve mental energy or breaking the cycle of ‘outsourced judgment’, we must resist the temptation to let AI displace our human duties of interpretation and care. By remaining ‘over the loop’, leaders ensure that technology remains a valuable aid rather than a force that blurs accountability. By choosing which cycles to feed and which to break, we transform our thinking from a closed circle into an upward spiral. Because in the architecture of thought, to understand where we are going, we must first look at where we have been.

Action Summary

The Pivot: Use the transition from single-loop to double-loop learning as the primary mechanism for shifting from a “bad” (stagnant) loop to a “good” (evolutionary) one.
For Good Loops: Implement active experimentation and reflective observation to ensure the cycle produces growth.
For Bad Loops: Identify recursive patterns that lack a “base case” or exit condition to prevent mental “stack overflow”.

Further Information

Readers interested in obtaining additional information on various concepts and models mentioned above may find the following sources helpful:

Systems Thinking (Peter Senge):
The Academy for Systems Change (Focuses on Senge’s work in organizational learning).
Systems Thinking (Donella Meadows):
The Donella Meadows Project (The official archive for her work, offering deep insights into the “leverage points” within a system and the core principles of Thinking in Systems).
Double-Loop Learning (Chris Argyris):
Harvard Business Review: Teaching Smart People How to Learn (The seminal article by Argyris on single and double-loop thinking).
The PDCA Loop (W. Edwards Deming):
The W. Edwards Deming Institute (The authoritative source for the Plan-Do-Study-Act/Check cycle).
The OODA Loop (John Boyd):
The Boyd Archive (Marine Corps University): (Search for John Boyd’s “A Discourse on Winning and Losing”).
Experiential Learning (David Kolb):
Experience Based Learning Systems (EBLS) (The official site for Kolb’s learning style inventory and cycle).
Reflective Practice (Graham Gibbs):
The University of Edinburgh – Reflection Toolkit: (A highly regarded academic summary of Gibbs’ 6-stage cycle).
AI Governance (Human in/over the loop):
NIST AI Risk Management Framework: (An authoritative framework for responsible AI use in organizational governance).

Disclosure: While NotebookLM was used to generate images and for editorial review, this article was conceived and written by a human being (both ‘in the loop’ and ‘over the loop’).