When Accountability Arrives Too Late
The Directive™
Why Institutions Must Detect Failure Before Harm Becomes the Evidence
By Samantha Avril-Andreassen, LLB (Hons), LLM, LPC, FRSA
Founder, SAFECHAIN™
Introduction
There is something deeply uncomfortable about the way institutions often learn.
A person is harmed.
A safeguarding failure occurs.
A serious incident is investigated.
A complaint escalates.
A judgment exposes deficiencies.
A regulator intervenes.
A public inquiry begins.
Only then does the system ask:
What went wrong?
Recommendations follow.
Policies change.
Training is commissioned.
Procedures are rewritten.
Assurances are given that lessons have been learned.
But there is a fundamental governance problem with this model.
The harm has already happened.
If institutional learning depends upon people being harmed before weaknesses become visible, we have created systems that are capable of responding to failure—but not necessarily capable of preventing it.
That distinction matters.
Because accountability after harm is necessary.
Governance before harm is better.
The Retrospective Accountability Model
Much of modern accountability is retrospective.
We investigate what happened.
We reconstruct timelines.
We examine decisions.
We identify missed opportunities.
We determine whether procedures were followed.
We make recommendations.
There is enormous value in this work.
Institutions must be accountable for their decisions.
But retrospective accountability has an unavoidable limitation:
It cannot undo the harm that revealed the failure.
A serious case review can identify missed safeguarding opportunities.
An inquiry can expose fragmented decision-making.
A regulator can identify inadequate governance.
A court can scrutinise evidence.
But by the time those mechanisms are activated, the institutional weakness may already have produced consequences.
The challenge for modern governance is therefore not simply becoming better at investigating failure.
It is becoming better at recognising the conditions in which failure is developing.
Harm Should Not Be the Diagnostic Tool
This leads to a difficult question.
How much harm must occur before an institution accepts that something is wrong?
One complaint?
Ten?
A regulatory intervention?
A judicial finding?
A safeguarding tragedy?
A public inquiry?
There should never be an institutional threshold at which human suffering becomes the evidence necessary to prove that governance was inadequate.
Yet systems can unintentionally operate in precisely this way.
Warning signs are treated individually.
Complaints are processed separately.
Professional concerns remain within departmental boundaries.
Near misses are resolved without wider analysis.
Patterns remain invisible because every event is examined as a standalone matter.
Eventually, something serious happens.
Suddenly, the pattern becomes obvious.
But the pattern did not begin with the final event.
The final event merely made it impossible to ignore.
The Warning Signs Usually Come First
Major institutional failures rarely emerge without warning.
The warnings may be subtle.
Repeated complaints.
Unusual case patterns.
Staff concerns.
Data anomalies.
Near misses.
Incomplete records.
Recurring procedural exceptions.
Delayed decisions.
Escalating safeguarding concerns.
Poor communication between agencies.
None may independently prove systemic failure.
That is precisely why governance matters.
Governance exists to recognise significance before certainty exists.
If institutions wait until the evidence of failure becomes overwhelming, they are no longer preventing risk.
They are documenting its consequences.
Complaints Are Governance Intelligence
One of the most underused sources of institutional intelligence is the complaint.
Complaints are frequently treated as administrative problems requiring resolution.
Was the complaint upheld?
Was the response sent within the required timeframe?
Was the complainant satisfied?
Was the matter closed?
Those questions matter.
But governance should ask something else:
What is this complaint telling us about the system?
A single complaint may identify an individual problem.
Repeated complaints concerning the same process may identify something much larger.
Complaints can reveal:
weaknesses in policy;
gaps between policy and practice;
inaccessible procedures;
recurring communication failures;
safeguarding blind spots;
inconsistent decision-making;
institutional fragmentation.
A mature organisation does not simply close complaints.
It mines them for intelligence.
Near Misses Matter Too
High-risk industries understand the importance of near misses.
An incident does not need to produce catastrophe before it becomes worthy of investigation.
The fact that catastrophe almost occurred is itself valuable information.
Public administration should embrace the same principle.
Where a safeguarding failure was narrowly avoided, ask why.
Where incorrect information was identified before a decision was made, ask how it entered the system.
Where a vulnerable person eventually received support only after repeated attempts, examine the earlier failures.
Where information was discovered by chance rather than design, investigate the governance gap.
The question should never be:
"Did anything terrible happen?"
It should be:
"What prevented something terrible from happening—and can we rely upon that protection next time?"
Evidence Integrity™ Requires Early Detection
Evidence Integrity™ is not simply about preserving evidence after something has gone wrong.
It must also enable institutions to detect emerging risk.
That means connecting information.
Identifying inconsistencies.
Recognising missing records.
Testing assumptions.
Examining patterns.
And understanding when the absence of information is itself significant.
A governance system that stores enormous quantities of information but cannot recognise emerging patterns does not possess intelligence.
It possesses data.
The distinction is critical.
Participation Integrity™ Is an Early-Warning System
People affected by institutional decisions often recognise system weaknesses before leadership does.
Service users experience barriers directly.
Families recognise recurring communication failures.
Survivors understand where safeguarding processes fail to reflect lived reality.
Frontline professionals see where policy and practice diverge.
Yet institutions can treat these perspectives as anecdotal rather than diagnostic.
Participation Integrity™ challenges that approach.
Meaningful participation is not simply about consultation.
It is a governance mechanism.
Listening to those experiencing the system can reveal weaknesses long before those weaknesses appear in formal performance indicators.
Institutional Fragmentation Hides Emerging Failure
Modern safeguarding rarely belongs to one organisation.
A person may simultaneously interact with healthcare, housing, police, social care, education, courts, benefits agencies, charities and financial institutions.
Each organisation may hold one fragment of the picture.
Each may correctly conclude that the information it possesses does not demonstrate significant risk.
But collectively, the information may tell an entirely different story.
This is one of the greatest challenges in modern governance.
Fragmented institutions produce fragmented truth.
And fragmented truth delays intervention.
The question therefore cannot simply be:
"What does our organisation know?"
It must increasingly become:
"What might the system know collectively that none of us can currently see individually?"
The Cumulative Harm Model™
Cumulative harm fundamentally challenges incident-based governance.
Significant harm may emerge through accumulation.
One delay.
One missed disclosure.
One administrative error.
One unanswered concern.
One fragmented decision.
One lost opportunity to intervene.
Each may appear manageable.
Together, they may transform someone's life.
The Cumulative Harm Model™ therefore requires institutions to recognise that severity cannot always be measured by examining incidents independently.
Sometimes the most important evidence lies in the relationship between them.
From Reactive Accountability to Preventive Governance
This is where governance must evolve.
Traditional accountability asks:
What went wrong?
Preventive governance asks:
What is beginning to go wrong?
Traditional accountability asks:
Who was responsible?
Preventive governance asks:
What conditions are making failure more likely?
Traditional accountability asks:
Were procedures followed?
Preventive governance asks:
Are those procedures actually controlling the risk they were designed to address?
Traditional accountability examines the past.
Preventive governance uses the past to protect the future.
Both are necessary.
But they serve fundamentally different purposes.
Leadership Must Ask for Weak Signals
Boards and senior leaders understandably want assurance.
They want dashboards.
Performance indicators.
Compliance reports.
Risk ratings.
But effective leadership cannot rely exclusively upon information designed to reassure.
Leaders must actively request information capable of challenging reassurance.
What are the recurring complaints?
Where are the near misses?
Which procedures repeatedly require exceptions?
Where are professionals expressing concern?
Which safeguarding risks are increasing?
Where is data incomplete?
What assumptions underpin our confidence?
What would we wish we had known if something went seriously wrong tomorrow?
That last question should appear in every boardroom.
The SAFECHAIN™ Perspective
SAFECHAIN™ approaches governance as an architecture of prevention.
Evidence Integrity™ protects the reliability and completeness of information.
Participation Integrity™ ensures that lived experience and affected voices inform institutional understanding.
Disclosure Integrity™ supports transparent and informed decision-making.
Process Integrity™ tests whether procedure continues to achieve its intended purpose.
Institutional Fragmentation™ identifies where disconnected systems conceal cumulative risk.
The Cumulative Harm Model™ examines how apparently minor failures can accumulate into significant harm.
Institutional Capability™ moves the question from whether an organisation complied to whether it was actually capable of achieving the outcome required.
Together, these principles point towards a different model of governance.
One in which institutions do not wait for catastrophe to tell them that something is wrong.
Accountability Must Become Predictive
No governance system can prevent every failure.
No institution can eliminate every risk.
That should never be the standard.
The standard should be whether institutions possess reasonable mechanisms for recognising emerging weaknesses, responding to credible warning signs and learning before harm becomes entrenched.
That is the difference between an organisation that merely possesses governance structures and one that possesses governance capability.
The goal is not prediction with certainty.
It is earlier recognition.
Earlier challenge.
Earlier intervention.
Earlier learning.
Because every stage at which failure can be identified before serious harm occurs represents an opportunity to protect someone.
Conclusion
Accountability matters.
Investigations matter.
Public inquiries matter.
Regulation matters.
Judicial scrutiny matters.
Lessons after failure matter.
But we should never confuse learning after harm with preventing harm.
The strongest institutions will be those capable of identifying weaknesses while those weaknesses are still small enough to correct.
Before the complaint becomes a crisis.
Before the near miss becomes a tragedy.
Before fragmented information becomes systemic blindness.
Before institutional drift becomes institutional failure.
And before another person has to become the evidence that proves the system was not working.
That is the next frontier of governance.
Not simply accountability after failure.
Accountability before failure becomes inevitable.
The Directive™
"The true measure of governance is not how thoroughly an institution investigates harm after it occurs, but how effectively it recognises the conditions that could allow that harm to occur in the first place."
Copyright Notice
© 2026 Samantha Avril-Andreassen. All Rights Reserved.
Published by SAFECHAINN Ltd (Company No. 12038453).
This publication forms part of The Directive™, SAFECHAIN™'s thought leadership and governance architecture examining safeguarding, institutional accountability, evidence integrity, participation, organisational capability and systems reform.
All original written content, analysis, concepts, methodologies and SAFECHAIN™ frameworks referenced within this publication are the intellectual property of Samantha Avril-Andreassen and SAFECHAINN Ltd, including The Directive™, Evidence Integrity™, Participation Integrity™, Disclosure Integrity™, Process Integrity™, Institutional Fragmentation™, The Cumulative Harm Model™, and Institutional Capability™.
No part of this publication may be reproduced, adapted, distributed, republished, stored or transmitted in any form without prior written permission, except for brief quotations for academic, educational, professional review or responsible journalistic purposes with appropriate attribution.
This publication is provided for research, education, policy and governance discussion. It does not constitute legal advice and should not be interpreted as commentary upon any individual person, organisation or specific legal proceedings.