GOVERNANCE
Accountability That Works: Building Mechanisms People Trust and Leaders Can Defend
By Dr Zamda Mutamuliza· 8 June 2026· 8 min read
In Brief
- Most organisations have accountability policies. Very few have mechanisms that make accountability real when it matters most.
- Accountability fails for predictable reasons: responsibility is diffused, escalation stalls, reporting channels are distrusted, senior leaders are treated as exceptions, and remedy arrives too late.
- Leaders should ask not whether an accountability system exists, but whether it would hold under pressure and whether they would know the answer before a crisis, not because of one.
Accountability is one of the most frequently invoked and least operationalised ideas in organisational life. Boards endorse it. Policies promise it. But when harm occurs or a significant decision fails under scrutiny, many organisations discover that accountability was never truly built into the system. It was assumed, narrated, or delegated, but not designed.
That is the central mistake. Accountability is not a virtue that appears because leadership values it. It is an institutional outcome produced by structure, incentives, visibility, escalation, and consequences. The UN Guiding Principles on Business and Human Rights make it clear that access to remedy is not peripheral to responsible practice. It is an integral part of the its very architecture. A mechanism that only processes cases is an administrative system. A mechanism that changes behaviour, imposes consequence, and enables remedy is an accountability system.Everything that follows rests on this distinction.
Mechanisms matter more than intent
Organisations often treat accountability as mainly a question of culture or leadership tone. Both matter, but tone without mechanism is fragile.
When a concern arises, what determines the outcome is rarely the values statement. It is the mechanism: who receives the concern, whether anonymity is possible, whether retaliation is prevented, whether evidence is reviewed independently, whether timelines exist, and whether the process can produce consequence or remedy. The OHCHR Interpretive Guide clearly states that organisations must not only avoid causing harm but also address it when it occurs. An organisation cannot plausibly claim to fulfil its responsibilities if it lacks a reliable mechanism to identify and address problems.
The first test: can people use the system safely?
No accountability mechanism works if people believe using it will harm them.
Reporting channels exist in most organisations. But people do not trust them. They fear retaliation, career cost, or the futility of raising a concern that will go nowhere. The CCAB guidance on speak-up culture could not be clearer on this: organisations must actively reduce the perceived risk of speaking up, rather than simply providing channels for doing so, because issues addressed earlier rarely develop into serious failures. The distinction between these two is important because opening a channel is a communication decision, whereas reducing the perceived risk of using it is a governance issue.
A mechanism that exists on paper but is avoided in practice is not a functional mechanism, and its usage without trust is compliance theatre.
The second test: does the mechanism reach power?
Many accountability systems fail not because no one reports concerns, but because the concern cannot travel far enough to matter.
It stalls in middle management. It is redirected into HR when the issue is governance. It is softened in tone as it moves upward, noted and contained without reaching those with authority to act. That is why escalation design matters. The OECD Due Diligence Guidance treats grievance mechanisms as part of a broader due diligence system, not an isolated add-on, precisely because surfacing issues is not enough. The question is not whether a channel exists. It is whether that channel can reach the level at which the underlying problem can actually be addressed.
Independence is not optional
The closer a mechanism sits to organisational power, the more important independence becomes. It does not always require a fully external process, but it requires enough structural separation that the people assessing a concern are not captured by the interests or authority of those being scrutinised.
Without that separation, well-designed systems become vulnerable: selective interpretation, procedural delay, informal influence, or reputational protection masquerading as fairness. The UNGPs describe credible mechanisms as legitimate, accessible, predictable, equitable, rights-compatible, and transparent. People do not need a perfect system before they will trust it. They need to believe it is not designed to protect the institution from the truth.
Accountability requires consequence and remedy
One of the most common reasons accountability mechanisms fail is that they confuse process with outcome. A concern is received. An investigation begins. A report is drafted. But no meaningful consequence follows, no learning is embedded, and no remedy is offered to those affected. The organisation then treats the existence of the process as proof that accountability worked. It did not.
Accountability is complete not when the institution has documented a response, but when that response meaningfully addresses the issue through sanction, restitution, structural reform, changed incentives, or public explanation.
Good accountability systems produce intelligence
There is a strategic reason accountability matters beyond correction: it generates institutional intelligence.
A well-functioning mechanism tells leadership where pressure is building, where incentives are distorting conduct, where policies are failing in practice, and where particular groups are experiencing the organisation differently. The UK Corporate Governance Code 2024 expects boards to assess whether desired culture is genuinely embedded, an expectation that becomes meaningful only when boards use accountability data as evidence, not only as assurance.
When complaints disappear entirely, wise leaders do not celebrate. They ask whether trust has increased, or whether silence has.
Global Standards Brief
Instrument | Jurisdiction | Relevance to Accountability |
Global | Establishes access to remedy as a core pillar of responsible practice, not a supplementary obligation. | |
Global | Operationalises the responsibility to address adverse impacts where they occur. | |
Global / OECD members | Requires tracking, communication, and remediation across operations and business relationships. | |
United Kingdom | Boards must monitor whether purpose and values hold in practice across the workforce. |
Intelligence Note: AI and the accountability gap in automated decision-making
Automated systems are creating accountability failures that existing mechanisms were not designed to handle. When an algorithm determines a disciplinary outcome, flags a worker for performance management, or screens a complaint for triage, the decision logic is typically invisible to the person affected, to the manager nominally responsible, and often to the governance function overseeing the process. That opacity does not eliminate accountability. It relocates it, usually to a place where no one is looking.
Organisations deploying AI in processes that affect people’s employment, access to services, or treatment within an institution should be able to answer three questions their current accountability frameworks rarely ask: who is responsible when the system produces a harmful output, how does an affected person contest a decision they cannot see, and what remediation is available when the harm is caused at scale rather than individually. The EU AI Act creates obligations around transparency and human oversight for high-risk AI systems, but those obligations only have force if the accountability mechanisms behind them are real.
The Accountability Readiness Diagnostic
Before your next board meeting, audit review, or governance reset, ask:
- Can the people most affected by our decisions actually access our accountability mechanisms?
- Do they trust those mechanisms enough to use them?
- Can concerns involving senior leadership or structural issues escalate independently?
- What consequences follow when wrongdoing or harmful decisions are substantiated?
- Can our systems produce remedy, not only record-keeping?
- What patterns are we learning from complaints, grievances, and speak-up data?
- If a serious case became public tomorrow, could we demonstrate the mechanism was credible before the crisis began?
If those questions expose uncertainty, the organisation does not yet have accountability that works. It has procedures that may or may not hold when tested.
Closing Reflection
Accountability mechanisms are most needed in exactly the moments they are most likely to fail: when the issue is senior, when the commercial stakes are high, and when the institution has most to lose from the truth surfacing. Organisations that build credible mechanisms before those moments arrive protect themselves in ways that no policy document can. Organisations that discover the gap during a crisis rarely close it in time to matter.
The question worth sitting with is not whether your organisation has an accountability system. It is whether that system would hold, and whether you would know the answer before the pressure arrived, not because of it.
GRIA Review publishes analysis on governance, human rights, responsible business, and institutional accountability. If this piece raised questions relevant to your organisation, explore our other articles or write for us.