As IT Security Auditors and Ethical Hackers, our job is to collect, analyze, and report security issues for a wide variety of client organizations. Most of the time this is a smooth process. Two recent engagements, however, stood out. Neither was a technical failure, and neither was caused by an attacker. Both left the client less secure than they should have been, and both were avoidable.
After performing this process roughly 2,000 times for hundreds of clients, we have seen nearly everything. This post walks through two situations we ran into, why they matter to your security, and what we (and you) can do differently next time.
Situation one: the finding that disappeared
During an IT security audit, our team identified a misconfigured wireless access point. The device broadcast two networks: a private production SSID for employees and an open guest SSID for visitors. The guest SSID had been connected to the production network. Anyone who joined that access point as a guest was placed directly onto the private network.
This is a serious issue. A non-employee, including someone sitting in a car in the parking lot, could reach systems that should only be available to trusted staff. Our team documented the issue and confirmed it multiple times, using different auditors on different devices.
When we presented results to the client's IT manager, we were told the issue was not possible. The wireless network was a managed environment, he explained, and every access point received the same configuration. We agreed that was the design, but this particular access point did not match it.
The IT manager asked us to stop the presentation and return to the site to test again. We ended the meeting and sent two security consultants to the building, about 30 minutes away. When they ran a wireless scan from the parking lot, the issue was gone. The results we had observed multiple times the day before could no longer be reproduced. We removed the finding from the report and presented revised results later that day.
We later learned what had actually happened. The IT manager had changed the access point configuration himself, without anyone else in the organization knowing, while continuing to insist that we had reported a security issue that did not exist.
Why this hurts the client's security
On the surface, this ended well: the hole was closed. In reality, the client lost far more than an hour of meeting time.
Leadership was given a false picture. The final report said the wireless network was sound. It was not. Executives, auditors, and insurers reading that report would believe a control worked when it had failed.
The person being audited controlled the evidence. The individual responsible for the configuration was also the only person who knew it had changed. When one person can quietly alter the systems under review and then dispute the results, the audit no longer provides independent assurance.
Exposure was never investigated. If guests could reach production systems, the next question is whether anyone did. With the finding removed, nobody had a reason to review DHCP leases, firewall logs, or authentication records for that period.
Nobody found the root cause. A "managed" environment produced an unmanaged result. Without an honest account of how the access point was misconfigured, the same mistake can happen again on another device, at another site.
The audit trail was broken. An undisclosed fix between discovery and reporting erases the evidence that an incident responder, compliance reviewer, or cyber insurance carrier would expect to see.
A finding that gets fixed during the audit is good news. A finding that gets erased from the audit is a hidden risk.
How we respond now
Documented proof for every critical finding. Since this engagement, we provide documented proof with findings of this type. That includes timestamped screenshots, the BSSID and MAC address of the specific access point, the IP address and subnet assigned to the test device, and evidence of the production hosts that were reachable. A finding backed by this record stands on its own, even if the condition is changed before the results meeting.
Report to an executive sponsor, not only to IT. Critical findings should go in writing, the same day they are confirmed, to a named leader outside the IT function as well as to the IT contact. When more than one person knows about an issue, it becomes much harder for it to quietly disappear.
Report remediated findings as remediated. A condition fixed during the audit is still a finding. The right outcome is listing it with a status of "Identified and remediated during the audit," along with the dates. That gives the client credit for the fix and keeps the record accurate.
Review the change records. Most wireless controllers and network platforms log every configuration change, including who made it and when. Requesting those logs for the audit window turns an unexplained overnight change into a documented event.
Recommend separation of duties. The report should recommend configuration change alerts that go to someone other than the administrator making the change, along with a root cause review of the misconfiguration and a log review for the exposure window.
Set ground rules for the presentation. A results meeting is for reviewing findings, not re-testing them. Disputed items are resolved afterward against the documented proof, without stopping the entire review.
Situation two: when a vendor does not play well
Part of every IT security audit is a full review of the client's cloud environment for misconfigurations, signs of unauthorized access, and alignment with CISA guidance and other recognized standards. To collect that detail, we ask for an administrative-level account to use during the audit and remove afterward. Our contract clearly states the level of access required, and we advise that every account created for our team uses MFA and is time restricted. Our process does not change anything. That is the first, and perhaps the most important, assurance a professional auditor makes.
An audit also does more than find security issues. It shows the client how well its vendors are actually performing. For that reason, clients do not alert their vendors ahead of time, and vendors are never part of the contract process or the audit kickoff. The goal is to see the environment as it really is, not as it looks after a vendor has had time to prepare.
As our team prepared to begin, the client explained that a third-party provider managed their cloud environment. The client had no internal IT staff and relied on several vendors to support their systems. The client contacted the vendor and explained that the audit had been mandated by their board of directors for compliance reasons. Several hours later, the vendor replied with a flat refusal: they do not provide administrative access to a customer environment for security reasons. We prepared a more detailed request, endorsed by the client, and sent it back. After another long silence, the request was denied again. We were frustrated, and our client was upset with their vendor.
Why this hurts the client's security
To be fair, a vendor's caution is not wrong on its own. Handing out administrator accounts on request is how environments get compromised. In this case, though, the safeguards a careful vendor should look for were already in place: the access was documented, approved by the client, protected by MFA, limited in time, and governed by a no-changes process. The refusal was not about protecting the client.
The client cannot see its own risk. The organization owns the data, the compliance obligation, and the liability, yet it has no independent view into how its most important systems are configured.
The vendor decides whether it gets reviewed. Third-party providers are one of the most common paths into mid-market organizations. A provider that can block an independent review of its own work is a risk that cannot be measured.
The board's mandate goes unmet. The audit report must now carry a scope limitation, and the board is left without the assurance it asked for.
The client does not truly control its environment. If the client cannot authorize access to its own systems without the vendor's permission, that dependence becomes a serious problem during an incident, a dispute, or a vendor change.
Why administrative access, and not read-only?
Vendors often ask why an auditor cannot simply use a read-only role. It is a fair question, and in theory it would be ideal. In practice, read-only roles in cloud platforms do not report consistently. Some settings are hidden from reader roles entirely, and others return incomplete or default values. The result is false positives: findings that look like misconfigurations but are not, or controls that appear missing when they are in place. That wastes the client's time chasing problems that do not exist and erodes trust in the findings that do.
An accurate audit requires an accurate view. Administrative access, used under a strict no-changes process, protected by MFA, and removed when the audit ends, is how we get one.
An audit that the vendor can veto is not an independent audit.
Better responses
The client should hold its own administrative access. This is the most important fix. Every organization should own its cloud tenant and keep at least one administrative account of its own, separate from its vendor. With that in place, the client can create a time-restricted, MFA-protected audit account directly, and the vendor's approval is never required.
Write the right to audit into the contract. At signing or renewal, vendor agreements should include a right to independent, unannounced security audits and a commitment to provision auditor access within a set timeframe. A vendor that has already agreed to these terms has no basis for a refusal.
Lead the request with the safeguards. When a vendor must be asked, the request should state up front that the access is administrative but governed by a no-changes process, protected by MFA, time restricted, and removed when the audit ends.
Explain why read-only falls short. A short, plain explanation of why reader roles produce inaccurate results and false positives answers the vendor's most likely objection before it is raised.
Put the client's authority in writing. A formal authorization letter, signed by an officer of the client and referencing the board mandate, makes clear that the environment belongs to the client and the client has approved the access.
Escalate to the right level. A support desk often defaults to "no." The client's account executive or the vendor's security leadership can approve requests that front-line staff cannot.
Treat the response as a finding. Because the audit also measures vendor performance, how a vendor responds is part of the result. Long delays, unexplained refusals, and a client that cannot reach its own systems all belong in the report.
Document the limitation honestly. If access is never granted, the report states it plainly: what was not reviewed, why, and what the client should do about it. Compensating evidence such as a vendor's SOC 2 Type II report can be noted, but it does not replace a direct review, and the board deserves to know the difference.
The common thread: independence
These two situations look different, but they share a root cause. In each case, the party whose work was being evaluated was in a position to control what the auditors could see. The IT manager quietly changed the configuration and then disputed the evidence. The vendor simply refused to allow the review. Either way, the client and its board ended up with less information than they paid for.
An independent audit only works when the evidence is preserved and access does not depend on the people being audited. Here is the short list we now confirm with every client before the engagement begins:
| Before the audit | What to put in place |
|---|---|
| Executive sponsor | A named leader outside the IT function who receives critical findings and attends the results presentation. |
| Documented proof | Every critical finding delivered with timestamped, device-specific evidence that stands on its own. |
| Change records | Access to change logs and controller audit logs for in-scope systems covering the audit window. |
| Client-held access | Confirmation that the client can grant administrative access to every in-scope system itself, without advance notice to the vendor. |
| Contract terms | Vendor agreements that include a right to independent, unannounced audits and timely access provisioning. |
| Reporting rule | Findings fixed during the audit stay in the report, marked as remediated, with dates. |
Final thoughts
An IT security audit is only as valuable as the accuracy of the report it produces. When findings disappear before they are documented, or when the people being audited can wall off key systems from review, the report looks better than reality. That is the most dangerous kind of report to hand a board of directors.
The good news is that both of these problems are solvable. Documented proof, an executive sponsor who hears about critical findings directly, and client-held access to every system in scope keep the audit independent, so it can do its job of finding risk rather than becoming a test to be passed. That is what Practical Security for the Mid-Market looks like in the field.
Planning an IT security audit?
Digital Elevation's Security Assessment Services include a pre-audit readiness call that covers rules of engagement, third-party vendor access, and critical findings handling, so your audit produces results you can act on and defend to your board. Start the conversation, or reach out directly at scott.montgomery@digitalelevationusa.com or 616-600-4737.
W. Scott Montgomery is Director of Security at Digital Elevation, where he leads the firm's security assessment and advisory practice. With more than 40 years of IT experience, Scott helps mid-market organizations, credit unions, and healthcare and defense-sector firms build practical, defensible security programs, from vulnerability management and examination readiness to compliance across frameworks including NIST CSF, FFIEC, GLBA, and HIPAA.