Skip to main content
COOKIE NOTICECOOKIE CHOICE

We use cookies for analytics (Google Analytics) and ad measurement (Google Ads, including remarketing, and ChatGPT Ads by OpenAI). Strictly necessary cookies keep the site working. See our Cookie Policy for what each one does and how to opt out.

We’d like to use cookies for analytics (Google Analytics) and ad measurement (Google Ads, including remarketing, and ChatGPT Ads by OpenAI), and none of them is set unless you accept. Cookie Policy

ISO 27001 · Inside the audit

What an ISO 27001 auditor actually asks

Saundhi ChauhanLead AuditorISO 27001 Lead Auditor · ISO 27701 Lead Auditor · ISO 42001 Lead Auditor

13 min readLast reviewed

The short answer

A Stage 2 auditor does not read your policies and tick boxes. They ask the people who do the work how they do it, and then ask for the records that would exist if the answer were true. So the questions are nearly always a pair: describe what you do, and show me the last three times you did it. The strongest answer in any ISO 27001 audit is a first-person description of a routine followed by the artefact that routine produces. The weakest is a correct quotation from a policy.

A finding at Stage 2 is a normal event, not a failure. What produces findings is a pattern across samples, not one nervous answer.

In this lesson

After this lesson you can:

  • Describe what happens in a Stage 2 audit, hour by hour, and why it is mostly interviews.
  • Prepare each role in your organisation for the questions that role will actually get.
  • Recognise the six answer shapes that create findings, and replace each with evidence.

Read first: The certification process and the mandatory documents

An audit is a conversation with evidence attached.

The mental model most people bring to a certification audit is an inspection. Someone arrives, reads the folder, and decides. That model survives Stage 1 and then collapses, because Stage 1 and Stage 2 are asking different questions.

Stage 1 is largely documentary. The auditor establishes whether your ISMS is designed, scoped and documented well enough to be audited at all, and whether the things that must have happened before certification have happened. Stage 2 tests operation, and you cannot test operation by reading. So the auditor samples: picks areas, picks people, picks time periods, and compares what people say to what the records show. ISO/IEC 17021-1 frames audit evidence this way, which is why no accredited audit skips the interviews.

Two consequences follow, and they are the whole reason this page exists. First, your documentation quality matters much less at Stage 2 than your records do. Second, the audit is distributed across your organisation rather than concentrated in a compliance function, and the people most exposed are the ones who have thought about it least.

Questions for leadership.

Senior management gets a short interview with a disproportionate effect on the auditor’s read of the whole system. The auditor is testing whether the ISMS has actual sponsorship or is a project someone was told to run.

  • What are the information security objectives for this year, and how do you know whether you are meeting them?
  • Walk me through the last management review. What was on the agenda, who attended, and what did you decide?
  • What is the most significant information security risk this organisation carries, and what are you doing about it?
  • What resources have you provided for the ISMS, and how were they decided?
  • Tell me about a time the ISMS changed a business decision.

The last one is the test. An executive who can describe a decision that went differently because of the ISMS has evidenced leadership in a way no policy signature can. An executive who cannot name a single objective has evidenced the opposite. Clause 9.3 sets out the inputs a management review must cover, and the auditor will have read your minutes against that list before walking in.

Questions for the ISMS owner.

This is the longest interview and the most technical. The auditor is establishing whether the management system has a functioning nervous system: risk in, controls out, findings fed back.

  • Show me your risk assessment. What method did you use, and who decided the acceptance criteria?
  • Pick a high risk from the register and walk me from the risk to the control to the evidence that the control operates.
  • Why is this control marked not applicable in your Statement of Applicability?
  • Show me your internal audit programme. Who audited this area, and how are they independent of it?
  • Take me through a nonconformity you raised yourselves. What was the root cause, and how did you verify the fix worked?
  • How did your scope get decided, and what changed in it this year?

Question three catches more organisations than any other in this list. Marking an Annex A control not applicable is entirely legitimate and it requires a justification that survives a follow-up question. “We do not do that” is not a justification when the auditor can see that you do. Question five is the one that distinguishes a live system from a built one, because an internal audit that found nothing tells the auditor where to look.

Questions for engineering.

If you build software, this is where the audit gets specific, and where a generalist auditor and a sector-fluent one diverge most visibly.

  • How does a change get from your laptop to production? Show me the last one.
  • Who reviewed that change, and could you have merged it without them?
  • How do you get access to production, and what happens when you no longer need it?
  • Where do secrets live, and how would I know if one had been committed to a repository?
  • What security testing runs before a release, and show me a result where it failed.
  • How do you know which third-party libraries you depend on, and what tells you when one has a vulnerability?

Notice how many of these ask for a negative case: a rejected change, a failed scan, a revoked access. An auditor asks for the negative case because a control that has never prevented anything is indistinguishable from a control that is not operating. Have one ready.

Questions for IT operations.

  • Show me the current asset inventory. How does a new device get onto it, and how does a disposed one come off?
  • Walk me through your last access review. Who was removed as a result?
  • What was the last patch cycle, and what is still outstanding and why?
  • Show me your backups, and show me the last time you restored from one.
  • Take me through your most recent security incident, including one you decided was not an incident.
  • What monitoring would tell you that something was wrong right now, and who sees it?

The backup question is the quiet one. Almost everybody backs up. A good number have never restored, and a backup nobody has restored from is a hypothesis.

Questions for HR.

HR is often the least prepared function and carries a surprising number of Annex A controls, from screening to terms of employment to what happens on the day someone leaves.

  • What screening is done before someone joins, and show me it happening for a recent hire.
  • Where in the contract are the information security responsibilities, and what survives termination?
  • Show me security awareness training records. Who has not completed it, and what happens then?
  • Walk me through a leaver. What was returned, what was revoked, and who confirmed it?
  • How does a role change get reflected in someone’s access?

The leaver question is the most reliable finding generator in the whole audit, because it crosses two functions. HR knows the person left. IT knows the accounts were disabled. Whether anyone can show those two facts connected, for the same person, on the same date, is a different matter.

What a weak answer sounds like.

None of these answers is dishonest. Each one simply fails to evidence operation, which is the only thing Stage 2 is measuring.

What was saidWhy it failsThe answer that works
"It is in the policy."Describes the rule, not the practice."We do X. Here is last month’s record of it."
"Our tool handles that automatically."No evidence the tool is configured as claimed.Open the tool and show the setting and its log.
"I would have to ask the compliance team."Suggests the control lives outside the work."This is my step. Here is where I record it."
"We have never had an incident."Reads as an untested process, not a safe one.Show the last test, drill or near miss you logged.
"That is not really how we do it any more."Document and practice have diverged.Either update the document or follow it. Before the audit.
Silence, then a glance at the consultantThe worst available answer."I am not sure, I will find out and come back." Then do.
Figure 1 Six answer shapes that create findings, and what to say insteadSource: ISO/IEC 17021-1 on audit evidence, and engagement experience

The pattern across all six is the same. Each replaces a specific past event with a general present claim. Auditors sample events, so give them events.

The three things that actually sink a Stage 2.

After a lot of audits, the serious problems cluster. None of them is about a weak answer in an interview.

One. No real internal audit before Stage 2. The standard requires you to audit yourself, and certification bodies expect it completed with corrective actions visibly in motion. An internal audit conducted two days before the certification visit, covering everything and finding nothing, is worse than none, because it tells the auditor the audit was performed for the certificate rather than for the organisation.

Two. A management review that did not cover the mandated inputs. Clause 9.3 specifies what has to be considered. A meeting happened, minutes exist, and half the required inputs were never discussed. This is a documentary finding that no amount of good practice elsewhere repairs.

Three. Documents and practice have diverged. Your procedure says quarterly and you do it twice a year. Your policy names a role that no longer exists. Your risk assessment predates the architecture. Each individually is minor, and together they establish that the documented system is not the operating system, which goes to the heart of conformity.

All three are discoverable before the audit, by you, which is the argument for a readiness pass rather than a hopeful Stage 2.

Preparing people without scripting them.

The instinct is to brief everyone on what to say. Resist it. A scripted organisation is obvious within two interviews, and the script itself becomes the thing the auditor investigates.

What works instead, in order of value:

  1. Tell each person which records are theirs, and where they live. Not the policy, the records.
  2. Have them open those records once, in advance, without help. The rehearsal is the retrieval.
  3. Give each person one negative example from their own area: a rejected request, a failed check, a revoked access.
  4. Explain that “I do not know, I will find out” is a good answer and a guess is not.
  5. Tell them a finding is normal and nobody is in trouble. Frightened people over-claim, and over-claiming is what gets written up.

Item five is not a morale measure. It is a control. The single most expensive thing that happens in audit interviews is a person inventing a process they think they are supposed to have, which the auditor then asks to see.

Would your team answer these today?

A readiness pass runs these questions against your actual records, role by role, and gives you the findings before the certification body does. That is the whole point of doing it in that order.

Key takeaways.

  • Stage 1 reads design and Stage 2 tests operation, which is why Stage 2 is mostly interviews with evidence attached.
  • Nearly every question is a pair: describe what you do, then show the last few times you did it. Answer in the first person and produce the artefact.
  • Have a negative case ready in every area. A control that has never stopped anything looks identical to one that is not running.
  • Leadership is tested on whether it can name objectives and describe a decision the ISMS changed, not on whether it signed the policy.
  • The leaver question crosses HR and IT and is the most reliable finding generator in the audit. Be able to show both halves for the same person on the same date.
  • Explaining a not-applicable control in your Statement of Applicability needs a justification that survives one follow-up question.
  • The three things that sink a Stage 2 are a hollow internal audit, a management review missing mandated inputs, and documents that no longer match practice. All three are findable in advance.
  • Prepare the evidence, not the words. Scripted answers are visible immediately and invite deeper sampling.

Quick check.

  1. An auditor asks an engineer how production access is granted. Which answer is strongest?

    1. "Our access control policy covers it."
    2. "The platform team handles access."
    3. "Through a ticket that needs my manager’s approval. Here are last month’s, including one I rejected."
    4. "Only senior engineers have production access."
    Check the answer

    C. It describes the practice in the first person, and it produces records on the spot, including a negative case. The other three describe a rule, delegate to someone else, or state a condition without evidencing how it is enforced. Clause 9.2, internal audit

  2. What is an auditor testing when they ask the same question to three different people?

    1. Whether staff have memorised the policy
    2. Whether the process operates consistently in practice
    3. Whether management is honest
    4. Whether the documentation is well written
    Check the answer

    B. Consistency across independent accounts is evidence that a process is real. Divergent answers do not prove dishonesty, they indicate that what is documented is not what is operating, which is the thing an audit exists to detect. ISO/IEC 17021-1

  3. Spot the mistake. A company briefs every employee to answer "that is covered by our information security policy" to any auditor question.

    Check the answer

    It replaces evidence of operation with evidence of coaching.. Stage 2 tests whether controls operate. A uniform non-answer tells the auditor nothing about operation and quite a lot about preparation, so it invites deeper sampling in exactly the areas the script was meant to protect. ISO/IEC 27006-1:2024

Sources.

Changes since publication

The specific questions on this page are the shapes auditors use in practice, written from engagement experience. They are not quoted from ISO/IEC 27001, which is sold rather than published freely. The requirements they test are on our clause pages, which are linked at each point.

Last checked against ISO’s catalogue entries for the certification-body standards, and our own clause reference pages on 9 November 2026.

Questions people ask.

Will the auditor interview ordinary staff or only management?
Both, and the staff interviews carry more weight than people expect. Management can describe what is supposed to happen. Only the person doing the work can show that it does, and the auditor samples evidence against what that person says. Expect the auditor to ask for specific people by role once they have read your process documents.
How long does each interview take?
Usually somewhere between twenty minutes and an hour per area, depending on how much of the scope that person covers and how readily the evidence appears. An interview runs long mainly when the evidence has to be found rather than opened.
Can we have the consultant in the room during interviews?
Generally yes as an observer, and it is a bad idea to rely on it. An auditor is specifically testing whether the organisation runs its own management system, so answers that route through a consultant undermine the thing being evidenced. Prepare your people properly and let them answer.
What if someone genuinely does not know the answer?
Saying so and offering to find out is an acceptable answer and a far better one than a guess. Auditors are experienced at telling the difference between a person who does the work and does not know a policy reference, which is normal, and a person who does not do the work, which is a finding.
Does one bad interview fail the audit?
No. An auditor is building a picture from many samples, and a nervous or incomplete answer in one area gets probed further rather than written up immediately. What produces findings is a pattern: the same gap appearing in several places, or evidence that contradicts what the documents say.
Will the auditor ask for live system access?
Often, in the form of looking over someone’s shoulder while they open a console, a ticket or a log. This is normally screen sharing or sitting beside the operator rather than credentials being handed over, and you are entitled to control how it happens under your own access rules.
What is the single most common finding at Stage 2?
Evidence that a required activity happened, without evidence that it happened on the schedule your own documents set. Access reviews performed once rather than quarterly, a risk assessment dated eighteen months ago, a management review that covered some of the mandated inputs. The activity exists and the discipline around it does not.
Should we rehearse the answers?
Rehearse the evidence, not the words. Make sure each person knows which records belong to them, where those records are, and can open them without help. Scripted answers are easy to spot and they invite the auditor to dig into whatever the script is covering.

Get in touch

Book a free consultation or send us your requirements. We respond within 24 hours.

Send us your requirement.

01Your details
02What is it about?
ISO 27001
03Send
Asking about: ISO 27001