Back to Blog

After the Cyber Security Assessment: How to Prioritise and Verify Remediation

By Jamie PopePublished 8 min readCyber Security
Remediation cards separating an implemented change from a verified security fix.

A cyber security remediation plan turns assessment findings into decisions, controlled changes and evidence that the intended improvement has happened. It should explain what needs attention first, who owns the work, what happens while a fix is pending and how completion will be checked.

The report is an input. Reduced exposure is the outcome.

Our previous guide explored what an advanced cyber security assessment should test. Once those questions have been answered, leadership faces a different challenge: how to act on the findings without losing important work between teams, suppliers and change windows.

Separate the finding from the decision

Start by confirming what each finding actually says. Record the affected service, supporting evidence, scope and limitations. Separate demonstrated weaknesses from issues requiring further investigation.

Remove duplicates without losing coverage. Ten observations might share a root cause, but fixing one configuration does not automatically resolve every affected system.

A suspected false positive needs investigation and a recorded explanation, not silent deletion. An item outside the assessment scope may need a new owner or further review rather than a quick closure.

Keep the original observation linked to the decision and subsequent work. That makes it easier to explain why an action was taken and what it was intended to achieve.

Prioritise business exposure, not just the score

Technical severity helps describe a weakness. It does not, by itself, determine the order of every business action.

Ask which systems are exposed, what access is needed, whether exploitation is reported by authoritative sources, and what the affected service enables. Consider the information involved, dependencies and existing safeguards.

The NCSC's triage guidance highlights internet-facing services and potential impact when prioritising issues. It also distinguishes fixes, acknowledged issues and items still under investigation.

Use business context to improve decisions, not to rationalise indefinite delay. Keep routine updates moving through an established process rather than making every update wait for a new committee discussion.

Where active exploitation or an urgent vendor advisory changes the situation, use the organisation's emergency response arrangements. A normal maintenance window should not become an automatic reason to leave an urgent exposure untreated. Follow the relevant advisory and assess whether incident investigation is also required.

Build a working remediation register

A practical register should be short enough to use and detailed enough to hold people accountable.

FieldWhat to record
Finding referenceLink to the original evidence and affected population
Business consequenceService, information or activity at risk
Priority and rationaleWhy this needs attention now or later
Action ownerPerson responsible for moving the work forward
Decision ownerPerson authorised to approve remaining business risk
TreatmentFix, replacement, restriction, retirement or further investigation
DependenciesSupplier actions, approvals, testing and maintenance windows
Target and review datesPlanned completion and earlier reassessment triggers
Interim protectionTemporary measures, their limits and expiry
VerificationMethod, reviewer and evidence required for closure

This is a recommended operating template, not a prescribed certification form. Use an existing service-management system if it already supports the workflow. Avoid maintaining competing versions in email, spreadsheets and tickets without a clear source of truth.

Give shared work one accountable coordinator

A supplier may need to release an update. An internal application owner may need to test it. A provider may perform the deployment. A business manager may approve downtime.

These can be separate responsibilities, but the finding still needs someone to coordinate the whole outcome.

“Waiting for supplier” is a description, not a complete plan. Record what was requested, who is chasing it, the next review date and what happens if the dependency remains unresolved.

In a co-managed IT arrangement, agree the handover explicitly. Define who can make changes, who provides evidence, who checks the result and who escalates delays.

Do not confuse delivery ownership with authority to accept risk. An engineer should not become the default risk owner simply because the ticket is assigned to them.

Plan the change and the evidence together

Before work starts, define what success will look like.

For example, a privileged-access finding might require more than removing one account from one group. The agreed scope may include related assignments, access through another platform and confirmation that essential administration still works.

For a Microsoft 365 configuration finding, record the tenant and relevant objects, approved settings, expected coverage and permitted exceptions. Capture evidence through authorised administrative processes. Do not assume that a change in one policy reaches every user or connected service.

Agree testing, communications and rollback arrangements proportionate to the service. A rollback may restore availability but also restore the exposure, so it needs an escalation and treatment plan.

Use existing change controls without allowing process overhead to obscure the actual decision. Where urgency demands an emergency change, record the authority and complete the required follow-up.

Keep temporary protection visibly temporary

Sometimes the permanent fix cannot be implemented immediately. A supported workaround, access restriction or other safeguard may reduce exposure while the longer-term action is arranged.

Describe what that measure does and does not address. Give it an owner, a review point and a condition for replacement or removal.

Do not mark a finding permanently resolved simply because the workaround has been applied. Equally, do not leave a useful temporary safeguard undocumented because it is not the final answer.

The distinction matters for reporting. Leadership should be able to see which risks have been reduced, which fixes are verified and which dependencies remain open.

Verify the outcome before closing the finding

A deployment record shows that a change was attempted or completed. Verification asks whether the original weakness has been addressed within the relevant scope.

Depending on the finding, that could involve an authenticated rescan, a configuration check, an access test or a safely authorised retest. Select a method that answers the original question.

The NCSC's verification guidance calls for checking fixes made through reconfiguration or mitigation and monitoring temporary workarounds.

Record the date, method, coverage, result and remaining limitations. For material findings, consider review by someone other than the implementer where practical.

If the result is inconclusive, keep that visible. “Awaiting verification” is more informative than “closed” when the evidence is not yet available.

Make accepted risk a deliberate business decision

Not every finding will be removed immediately. There may be competing operational needs, technical constraints or a planned replacement.

That decision should have an authorised owner, a reason, an understanding of the remaining exposure and a review date. Check applicable requirements before treating acceptance as an available option.

Do not let silence, a missed deadline or an unavailable supplier become implied approval. Nor should a temporary acceptance become permanent because nobody revisited it.

Escalate when the situation changes: increased exposure, new exploitation information, a failed safeguard or a delayed replacement may invalidate the original reasoning.

Report progress without rewarding easy closures

A falling ticket count can hide the fact that the most consequential finding has not moved.

Report material risks reduced, verified outcomes, ageing high-priority items, blocked dependencies and decisions needed. Separate implemented changes from verified fixes and temporary mitigations.

For example, “Five changes implemented, three verified, two awaiting supplier validation” is more useful than “Five issues resolved” if the evidence supports only the first statement.

Avoid implying that any single completion percentage proves the organisation secure. Reports should explain scope and uncertainty as clearly as progress.

Start with the findings that need a decision

Choose a manageable group of important findings. Confirm the priority, owners and dependencies, then agree the change and verification work together.

Cyber security risk assessment support can help clarify priorities. Co-managed delivery can add agreed capacity and specialist support alongside the internal team. The engagement should state who does what and what evidence is expected.

Turn your assessment into an actionable improvement plan. Discuss remediation planning and co-managed delivery with Cloud Agile.

Frequently asked questions

What is a cyber security remediation plan?

It connects findings to prioritised treatment, owners, dates, dependencies and verification. It should also record temporary safeguards and unresolved risk decisions.

Should we fix every critical finding before every high finding?

Severity is important, but exposure, exploitation information and business consequences also matter. Investigate context promptly and document priorities rather than applying a score mechanically or using context to justify delay.

Who owns remediation when several suppliers are involved?

Allocate each task to the relevant party and appoint an accountable coordinator for the finding. Keep authority to accept business risk separate from technical implementation.

Is applying a patch enough to close a finding?

Not automatically. Confirm the intended systems received the change and use an appropriate check to establish whether the weakness has been addressed. Some situations also require investigation for prior compromise.

Can a temporary workaround count as remediation?

It can be a treatment measure, but its limits and temporary status should remain visible. Track the permanent action and review whether the workaround remains effective.

How long should remediation take?

Set targets based on risk, authoritative advisories and applicable obligations. Use an urgent route for time-critical exposure and review delays explicitly. No single deadline is appropriate for every finding.