An ISO 27001 internal audit is a check your own organization runs, at planned intervals, to confirm that your information security management system (ISMS) meets the standard and your own rules, and that it works in practice. The standard requires it, and certification does not happen without it. The audit goes well when it has a written programme, an auditor who did not build what they are testing, evidence for every control in scope and findings that end in tracked fixes. If you want an experienced team to run the audit and the fixes under one agreement, that is what our compliance assessments are for.
TL;DR
- Clause 9.2 of ISO/IEC 27001:2022 requires internal audits at planned intervals, run from a documented audit programme.
- Audit once a year at minimum for most organizations, and more often for high-risk areas or after major changes.
- Auditors must be competent and must not audit their own work. Small teams often need an outside party for that reason.
- Start with the Statement of Applicability, then sample evidence for every control you have selected.
- Every finding needs an owner, a root cause, a due date and proof that the fix worked.
Table of Contents
- What an ISO 27001 internal audit is
- What Clause 9.2 requires
- Who can run the audit
- The audit process, step by step
- An internal audit checklist that works
- Building an audit schedule
- What to do with findings before certification
- Common mistakes
- Where we fit
- Our perspective
- FAQ
- Sources
What an ISO 27001 internal audit is
Your ISMS is the set of policies, risk decisions, controls and routines you use to protect information. ISO 27001 is the international standard that describes what that system must contain. An internal audit is the first-party check of it: your organization, or someone acting for it, compares what you do against the standard and against your own documents.
It is different from the certification audit. A certification body (an independent auditing company) runs that one, and its result decides whether you receive the certificate. The internal audit comes first and has a different purpose: finding gaps while they are still cheap to fix and while only you see them. The certificate is then valid for three years, with surveillance audits by the external auditor in the years between, as the Cloud Security Alliance describes. The internal audit does not stop after the first certificate. It is a recurring part of the system.
What Clause 9.2 requires
The current edition is ISO/IEC 27001:2022, the third edition, published in October 2022. The 2013 edition is withdrawn. A 2024 amendment added climate-action wording, covered below.
In the text of the 2022 standard, Clause 9.2 has two parts:
- 9.2.1 General. You conduct internal audits at planned intervals to find out whether the ISMS conforms to your own requirements and to the standard, and whether it is effectively implemented and maintained.
- 9.2.2 Internal audit programme. You plan, establish, implement and maintain a programme that sets frequency, methods, responsibilities, planning requirements and reporting. It must take into account how important the processes are and what earlier audits found.
The word that matters is programme. One audit before certification does not satisfy it. A certification auditor will expect to see a plan for the whole cycle, the criteria and scope for each audit, the reports and what happened afterwards.
What changed in the 2022 edition
For an audit, the change that matters most is Annex A. The 2022 revision reduced it from 114 controls in 14 clauses to 93 controls in four themes: 37 organizational, 8 people, 14 physical and 34 technological. It added 11 new controls, merged 24 and updated 58.

The February 2024 amendment is small. Clause 4.1 now asks you to decide whether climate change is a relevant issue for your ISMS, and a note in 4.2 says interested parties may have climate-related requirements. Your auditor should see that you considered the question and recorded the answer. If you are still mapping the controls, our piece on why the 93 Annex A controls are not a checklist explains how to treat them as options driven by risk.
Who can run the audit
The auditor needs enough knowledge, skill and experience to assess an ISMS against the standard, and you should record why you judge them competent. Just as important, they must be impartial. Nobody should audit their own work or an area they are directly responsible for.
That rule decides who does the job in most organizations:
| Organization | Typical problem | Practical answer |
|---|---|---|
| Small business, one IT person | The person who built the controls is also the only one who understands them | Bring in an outside party for the audit |
| Startup with a security lead | The lead owns the ISMS and cannot audit it | Audit by someone outside the security team, or externally |
| Mid-size or enterprise | Internal audit staff may lack ISMS knowledge | Train an internal auditor or add specialist support |
| Healthcare provider | Security, compliance and IT overlap with clinical systems | Auditor who knows both the standard and the regulatory setting |
An outside auditor is fully acceptable, because the standard asks for an audit under your programme, not for your own employees. What it does not accept is an audit carried out by the people being audited.
The audit process, step by step
The steps below follow the structure the standard sets out in 9.2 and 10.2, with the places where audits usually go wrong.
1. Define scope, criteria and objectives
Write down what this particular audit covers: which parts of the ISMS, which sites, systems and teams, which clauses and Annex A controls, and the period of records you will examine. The criteria are the standard plus your own policies and any contractual or legal requirements you have put into the ISMS.
Where it goes wrong: a scope that is wider than the one in your certification scope statement, or narrower, so that parts of the ISMS are never looked at. Over a full cycle, every clause and every selected control should be covered at least once.
2. Prepare an audit plan
Fix the dates, name the auditor, list who will be interviewed and agree the timeline. Share the plan in advance, so process owners can bring the right records instead of improvising.
3. Review documents first
Before any interview, read the documented ISMS. The core set is the scope statement, the Statement of Applicability (the list of controls you have selected, with the reasons), the information security policy, the risk assessment and risk treatment plan, management review minutes, corrective action records and your business continuity policy.
Start with the Statement of Applicability. If it still follows the 2013 Annex A list instead of the 93 controls of 2022, nothing else in the audit will line up with it. That is the first gap to close.
4. Interview, observe and sample
Talk to the people who run each control. Ask them to show you the process, not describe it. Then sample records: for every control you have selected, look at at least one real piece of evidence, such as an access review, a restored backup, a signed approval or a closed incident ticket.
Where it goes wrong: auditing the policy instead of the practice. A password policy that exists is a document. A system that enforces it, and a log that proves it, is a control.
5. Evaluate and classify findings
For each gap, state the requirement, what you found and the evidence. Most organizations sort findings into three kinds:
- Major nonconformity: a breakdown that affects the ISMS's ability to achieve its intended results.
- Minor nonconformity: an isolated gap that does not undermine the system as a whole.
- Observation or opportunity for improvement: not a nonconformity, but worth acting on.
The standard does not set universal thresholds between major and minor. Write your own definitions into the audit programme so findings are classified the same way every time, and so they match what a certification body is likely to apply.
6. Close the audit and report
Hold a closing meeting with the people audited, so nothing in the report surprises them. Then send the written report to the relevant management. Management review under Clause 9.3 is where those results become decisions on resources and priorities, so the audit report is an input to it.
7. Fix and verify
Clause 10.2 deals with nonconformity and corrective action. Each action needs an accountable owner, a root-cause analysis, a defined fix with a due date and retained evidence of completion. The last step is the one most often skipped: verifying that the fix addressed the cause and keeps working. A ticket marked closed is not a verified fix.
An internal audit checklist that works
A good checklist is a map of evidence, not a list of yes-or-no boxes. Cover Clauses 4 to 10 of the management system and the Annex A controls you selected. This table shows what to ask for in each area.
| Area | Evidence to review |
|---|---|
| Clauses 4 to 10 (context, leadership, planning, support, operation, evaluation, improvement) | Policies, risk assessment, management review minutes, improvement records |
| Risk management | Risk register, risk treatment plan, Statement of Applicability, control evidence |
| Annex A controls | Access logs, encryption configurations, backup and restore records, incident records |
| Documentation | Policy versions, approval records, document control logs |
| Awareness and training | Training records, awareness session logs, staff interviews |
| Incident management | Incident logs, response records, post-incident reviews |
Two habits make a checklist stronger. First, trace one item end to end: pick a risk from the register, follow it to its treatment, then to the control that carries it out and the evidence that the control ran. Second, ask staff, not just managers. If an engineer cannot describe the incident process in the policy, the policy is not working.
Technical controls produce the clearest evidence. Patching is a common example, and a written policy backed by a patch record is easy to audit; our guide to an audit-ready patch management policy shows what that evidence looks like. Incident handling is another: a incident response checklist gives you the records an auditor will ask to see.
Building an audit schedule
The standard says "planned intervals" and leaves the length to you. The Cloud Security Alliance notes that you set the interval according to size, complexity, legal, regulatory and contractual requirements, changes, risk assessments and earlier audit results, and that at least once a year is reasonable.
A schedule that holds up in practice:
- Set a full-cycle plan. Spread the audit over the year, or cover every clause and control over the three-year certification cycle, with higher-risk areas more often.
- Put the audit before the external one. Complete it, and the corrective actions, early enough that the certification or surveillance auditor sees closed findings, not open ones.
- Add triggers. A new cloud platform, an acquisition, a serious incident or a change of scope justifies an extra audit.
- Weight by risk. Areas that failed last time, or that carry the most sensitive data, get audited earlier and deeper.
- Book the auditor and the interviewees at the start of the year. Audits slip when they depend on the calendar of one overloaded person.
For budgeting the whole certification effort, including the audits and the external fees, see our breakdown of US ISO 27001 certification costs. For the order of work from the start, the implementation steps for US teams show where the first internal audit sits.
What to do with findings before certification
Treat the internal audit report as your pre-certification punch list. Work through it in this order:
- Fix anything that looks like a major nonconformity first, since that is what stops a certificate.
- Make sure every finding has an owner and a date, and that the evidence of the fix is filed where an auditor can find it.
- Update the risk register and the Statement of Applicability if findings changed your risk picture.
- Take the report to management review and record the decisions that follow.
A certification auditor is not alarmed by findings in an internal report. They are reassured by findings that were found, owned and closed. The problem is a report with no findings at all, which suggests the audit was not deep enough.
Common mistakes
- An auditor who built the system. This breaks the impartiality requirement.
- Checking documents only. The standard asks whether the ISMS is effectively implemented, which means looking at operation.
- No programme. A single audit with no schedule, criteria or reporting rules does not meet 9.2.2.
- Closing findings without verification. Clause 10.2 expects proof that the fix worked.
- Out-of-date Annex A. Using the 14-clause structure from 2013 is a quick way to attract a major finding.
Where we fit
If you run security with a small team, or a regulated environment such as healthcare, the hard part of the internal audit is rarely understanding the standard. It is finding an auditor who is independent of the work, and then finding the time to fix what they report. We provide compliance assessments for ISO 27001 as well as SOC 2, HIPAA and PCI-DSS, alongside managed IT and cybersecurity services under one agreement with one point of accountability. If you also plan to certify against SOC 2, our guides on a readiness assessment and a gap analysis show how much of the evidence overlaps.
See our compliance and risk services to set up an assessment.

Fixed-fee, flat monthly pricing is how we work, so you can plan the audit and the follow-up work against a predictable budget.
Our perspective
Most failed audits we see described are not failures of knowledge. They are failures of ownership. The security lead writes the policy, runs the control and then audits it, or the audit finds six gaps and nobody is named to fix them.
We built our approach around a single point of accountability. When compliance assessment, cybersecurity, cloud and IT management sit with one provider under one SLA, a finding from the audit goes to the same team that can change the configuration, retrain the staff or update the policy. That shortens the path from report to verified fix, which is the part of Clause 10.2 that organizations find hardest.
For a small business, that means an audit without hiring a full-time compliance function. For a startup, it means evidence that grows with the cloud platform. For a healthcare provider, it means security and compliance work that fit together. We would rather you walk into the certification audit with closed findings than with a clean-looking report.
Nick - Sr. Executive
FAQ
Does ISO 27001 require an internal audit?
Yes. Clause 9.2 requires you to conduct internal audits at planned intervals and to keep an audit programme. The certification body will ask to see the programme, the reports and the corrective actions.
How often should we run an internal audit?
The standard leaves the interval to you, based on your size, complexity, risk and past results. At least once a year is a reasonable baseline, with extra audits after major changes or incidents.
Can the person who manages our ISMS audit it?
No, not for their own work. Auditors must be competent and impartial, and cannot audit areas they are directly responsible for. An outside auditor is a common solution for small teams.
What should I review first?
Start with the Statement of Applicability and check that it reflects the 93 controls in four themes from the 2022 edition. Everything else in the audit builds on it.
What happens if the audit finds a major nonconformity?
Give it an owner, find the root cause, set a due date and fix it before the certification audit. Keep the completion evidence and verify that the fix lasts.
Sources
- ISO/IEC 27001:2022, information security management systems
- Full text of the ISO/IEC 27001:2022 standard
- ISO/IEC 27001:2022/Amd 1:2024, climate action changes
- ISO 27001:2022 Internal Auditor Training materials
- Cloud Security Alliance: values of ISO/IEC 27001 internal audit
