This is the fifth and final article in a series exploring each of the 5 Lenses, the framework I use to help boards and leaders see more of the system when navigating complex change. Last month, I explored the Process lens and what it means when the way work is designed distributes load in ways nobody intended and nobody sees. This month, I turn to the Practice lens, because even when the design is sound, there is a prior question: is the thing actually happening? Is it being sustained? And is it producing any learning at all?
This is not an article about compliance. It is about why organisations can have all the right structures and still not learn from them.
Take a charity with a data protection policy. It was written carefully, probably in 2018 when GDPR came in, by someone who took the responsibility seriously. It is on the intranet. Staff are pointed to it during induction. The organisation can produce it on request.
Look more closely and you will find it still references EU GDPR frameworks that apply differently in the UK post-Brexit. There is nothing in it about AI-processed personal data, because the policy predates the tools the organisation is now routinely using. The procedure it describes references a case management system that was replaced two years ago.
Nobody made a decision to leave it like this. It simply was not looked at.
That is not a knowledge failure. It is a Practice failure: the organisation learned in one place and continued doing in another, and never built the bridge between them. Nobody had sustained, protected time and clear ownership to make the connection. That is not a personal failing. It is a structural one. Saying "structural" does not mean nobody is responsible. It means the responsibility lies with whoever designed the conditions, not with whoever is working inside them.
The Practice lens does not ask whether you have a policy, a framework, or a training record. It asks whether the learning is actually happening in the work, whether it is being sustained, and whether it is changing anything.
One organisation I work with realised, during a routine team meeting, that three people had independently adapted the same process to work around the same problem, each without knowing the others had done it. None of them had raised it formally, because the adaptation felt too small to escalate. When the manager named it out loud, what followed was a twenty-minute conversation that changed how the team worked. The process was updated the following week.
The conditions for it were simple: a manager who asked, a team that felt safe enough to answer, and protected time to have the conversation rather than just get through the agenda. That is double-loop learning: not a structural intervention, but a moment where an assumption was questioned and changed as a result of what the work was showing.
One of the most useful distinctions I have found in this work is the difference between Process and Practice. They are related but they are not the same, and conflating them is how organisations end up convinced they are doing something they are not.
Process is the design. The structure, the documented intention, the agreed sequence of steps.
Practice is the actual doing. Whether the thing happens at all. How it is sustained over time, under real conditions, with real people who are busy and tired and carrying several other priorities. And critically, whether the organisation learns from doing it, adjusts, and does it differently as a result.
You can have a beautifully designed supervision framework and no consistent supervision in reality. You can have a data protection procedure and a gap between that document and Tuesday afternoon that is wide enough to drive a data breach through.
But the reverse is also true: you can have strong Practice with very little formal Process. Teams that have developed genuine habits of learning together, of changing how they do things based on what they find. That learning is real and valuable, even when it is not documented.
The organisational instinct when learning breaks down is almost always to fix the Process. Commission a new framework. Update the policy. But if the Practice was not happening inside the existing framework, a better framework will not fix it. The harder question is why the Practice degraded. And that question almost never has a simple answer.
NCVO's Governance in Focus report (2025), drawing on insights from over 800 trustees, staff and volunteers in around 70 organisations, offers a precise picture of what this looks like at governance level. Trustees rated their governance up to 15% higher than staff in the same organisations, particularly on culture, inclusion and leadership. Public summaries highlight that culture and behaviour are rarely reviewed systematically at board level; in many organisations, there is no regular practice of reviewing board culture at all.
Most trustees I work with are deeply committed. They give significant time, bring real expertise, and care about doing it well. The perception gap the NCVO data reveals is not about effort or intention. But a gap this consistent across organisation types and sizes is not simply a matter of different vantage points. It points to a power dynamic: the people who define what good governance looks like are the same people rating whether it is being achieved. The mechanisms by which what staff, volunteers, and frontline teams know actually reaches the board are not working as well as many boards believe they are - and boards, by definition, are the ones with the authority to change those mechanisms. The board is learning from its own view of the organisation. The staff and volunteers are living in it. Those are not the same source, and the gap between them does not close by accident.
A 2025 Harvard Business Review study adds a specific layer. Surveying 1,160 managers, it found that middle managers score lower on psychological safety than both their seniors and their direct reports. These are the people most responsible for transmitting what they see upward and translating decisions downward. They are also the least safe to do either honestly. When speaking truth about what is not working carries a cost, people stop speaking it. The Practice of surfacing reality degrades, gradually and invisibly.
The AI thread running through this series arrives here at its sharpest point. Early findings from the 2026 Charity Digital Skills Report survey show that 88% of charities now use AI tools in their day-to-day work, up from 76% in 2025. Yet interim analysis suggests that under half are using AI in a deliberate, strategic way, even though the proportion describing their use as "active" or "strategic" has almost doubled since 2025. That remaining gap is not a technology problem. It is a Practice problem: the reflective infrastructure that would turn individual action into organisational learning is not in place. The pattern is not unique to AI - it shows up in data protection, safeguarding, supervision, and governance review. AI makes it more visible because the stakes are higher and the pace of change is faster. But the underlying failure is the same one.
The Charity Excellence AI readiness data (April 2026) maps the same pattern in governance terms. Operational safeguards have improved, with several now rated Green. All three board-level governance controls remain Red: strategy, designated oversight, and ensuring staff are trained and compliant. Organisations are doing. They are not yet learning from the doing at the level that would change how they govern it.
One recent qualitative study on charity AI use captures this directly, with a respondent describing parts of their team as "trapped" by AI tools, feeling they could not do their jobs without them. That feeling may partly reflect the discomfort of a transition phase, the natural adjustment when new tools change established habits. But the pattern it points to is real: without a sustained habit of asking whether the tool is still serving the work, or whether the work has quietly started serving the tool, dependency becomes invisible until it is entrenched.
At board level
The Practice question is not whether the organisation has a governance review. It is whether the board changes what it believes as a result of what it learns. A board that scores its own culture significantly higher than the people who work inside that culture is unlikely to be receiving fully accurate information about itself. The most useful question a board can ask is not "do we have a process for this?" It is: when did we last change something we believed about this organisation, because of what we learned from the people doing the work?
At manager level
Practice either lives or dies. The team leader, service manager, programme lead: these are the people who see the gap between the designed process and the actual doing. The failure mode at this level is quiet compensation: running a version of the practice that looks like the framework but has become something else, shorter, shallower, safer. Not from negligence. From the entirely rational calculation that raising the gap costs more than absorbing it.
The better question for anyone in this layer is: what is the practice I am running that has drifted furthest from what it was designed to do, and does anyone above me know?
At individual and frontline level
Practice is lived experience. The classic failure in the charity sector is not that frontline workers and volunteers lack insight. They usually have more of it than anyone else in the system. It is that there is no reliable, sustained Practice by which that knowledge travels anywhere useful. The learning evaporates at the end of the shift, because the structure between the person who has it and the person who could act on it was never designed to carry it.
When Practice breaks down at frontline level, it lands hardest on people who have the least ability to go elsewhere or escalate - the people the organisation exists to serve. A support worker who knows a care process is not working, and has no route to change it, will keep delivering that process to the people who depend on it. Those people are at the end of a chain of undisclosed drift. The learning failure is not just an organisational problem. It is a quality and accountability problem, and the people bearing the cost of it are the ones with the least power to name it.
When organisations discover drift, the instinctive response is to ask the people closest to the work to fix it, on top of everything else they are already doing. The review needs an owner with the capacity to do it genuinely. If there is no such person, that is itself information from the Practice lens.
If you want to start somewhere this week, without a review process or a meeting to schedule:
Think of one practice your organisation treats as working. Not a policy, not a document. The actual doing: a supervision conversation, a team check-in, a reporting routine. Ask one person who runs it what it last changed. If they cannot answer, you have your starting point.
From there, three questions are worth sitting with before the week is out.
Pick one policy your organisation relies on. When was it last read, tested against current reality, and updated?
Pick one learning practice and ask not whether it happens, but what it last produced. When did it change something?
If a frontline worker or volunteer learned something important about what is or is not working for the people you serve, what is the actual route by which that learning reaches someone who can act on it?
If your organisation is using AI tools, when did someone last ask whether a specific tool is still doing what you thought it was doing, and who had the authority to change the answer?
If you want to go further, take a look at the 5 Lenses framework and the AI Governance Readiness Diagnostic, both are designed to structure exactly these conversations.
The Practice lens is the last in this series because it is the one that asks whether the other four are alive. Purpose tells you why the work matters. People tells you how it lands on the individuals doing and receiving it. Power tells you who can actually move the problem. Process tells you how the load is distributed. Practice asks whether the organisation is learning from all of that, genuinely, sustainably, in the actual doing of the work.
You can have clarity of purpose, a thoughtful approach to people, sound power distribution, and well-designed processes, and still have an organisation that does not learn from what it is doing. Because the practice of reflection has been scheduled but not protected. Valued in the strategy but not resourced in the week.
The Practice lens does not ask whether you have a learning framework. It asks when you last changed something you believed, because of what the work showed you.
Five lenses, one underlying question. Not whether you have the right framework, but whether the conditions you have built make the doing of it possible, for the people who do it, for the people who depend on it, and for the organisation that will need to keep doing it long after the strategy document has been filed.
Argyris, C. and Schön, D.A. (1978). *Organisational Learning: A Theory of Action Perspective.* Addison-Wesley.
Hagen, J.U. and Zhao, B. (2025). Middle managers feel the least psychological safety at work. *Harvard Business Review*, 22 October 2025. View article
NCVO (2025). *Governance in Focus: Charity governance is getting stronger, yet culture gaps persist.* NCVO. View report
Ramsay, N. (2026). Early insights from the 2026 Charity Digital Skills Report survey. Zoe Amar Digital, 26 March 2026. View article
McLintock, I. (2026). *From Quiet Use to Strategic Leadership: AI in the Charity Sector 2026.* Charity Excellence, April 2026.
*Note: Some figures cited above are drawn from interim analyses and early findings that may be updated as full reports publish.*
This article was co-created through a human-led process using several AI models as thinking partners. It reflects our commitment to ethical, transparent, and accountable use of AI, where human judgement, curiosity, and oversight remain central.