Security Questionnaires for SaaS: How to Answer Vulnerability Scanning Questions with Evidence
Written on . Posted in Informational.
Sooner or later, a bigger customer sends you a spreadsheet. It might be a CAIQ, a SIG, a HECVAT, or a homemade workbook from procurement. Somewhere in the middle sits a cluster of questions about vulnerability management: how often you scan, what you scan, who reviews the results, how fast you fix critical issues, and whether you can prove any of it.
For a small SaaS team, that cluster is where deals slow down. The founder knows the app is patched and a scanner runs "regularly," but the reviewer on the other side wants something firmer than good intentions. This guide explains what those questions are really asking, how to back each answer with scan evidence you already have (or can start producing this week), and how MSPs can do the same for the clients they look after.
Why vulnerability questions carry so much weight
Security questionnaires exist because your customer is taking on some of your risk. If your platform is breached, their data goes with it. Most sections of a questionnaire ask about policies, which a reviewer has to take on trust. Vulnerability management is different: it is one of the few areas where a vendor can show recent, dated, independent output from a tool, so reviewers tend to push harder there.
The widely used frameworks make this explicit. The Cloud Security Alliance's Cloud Controls Matrix and CAIQ v4 include a Threat and Vulnerability Management (TVM) domain with ten controls. In CSA's own implementation guidance for the TVM domain, the vulnerability identification control calls for procedures to detect vulnerabilities on organizationally managed assets at least monthly. Other TVM controls cover a remediation schedule with both scheduled and emergency responses, risk-based prioritization using an industry recognized framework, and tracking and reporting of identification and remediation activities.
The Shared Assessments SIG spans 19 risk domains and is updated at least annually, and in higher education the EDUCAUSE HECVAT plays a similar role for vendors selling to colleges and universities. The wording differs, but the vulnerability management questions are very similar across all of them.
What the questions are really asking
Strip away the framework labels and most vulnerability management questions fall into five themes. Each one maps to a specific piece of evidence.
| What the reviewer wants to know | Typical wording | Evidence that answers it |
|---|---|---|
| Do you scan at all, and how often? | "Do you perform vulnerability scans at least monthly?" | Scan schedule configuration plus a dated history of completed runs |
| What is in scope? | "Are all internet-facing systems and applications scanned?" | A target list that matches your production hosts, domains, and web apps |
| Is it more than a port check? | "Do you perform authenticated or application-layer testing?" | Network scan reports plus web application (DAST) reports, ideally authenticated |
| What happens to findings? | "What are your remediation timeframes by severity?" | A written SLA plus run-over-run results showing issues closing |
| Who decides on exceptions? | "How are accepted risks and false positives documented?" | Recorded decisions on individual findings with a reason and an owner |
Only one of these themes is about the scanner itself; the rest are about scope, follow-through, and record keeping. That is why "we use a scanner" rarely satisfies a careful reviewer, and why a modest but well documented program often scores better than an expensive tool nobody looks at.
Claims versus evidence
A claim is a sentence in a spreadsheet cell. Evidence is something the reviewer could check if they wanted to. Compare two answers to "Describe your vulnerability scanning program":
- Claim: "We regularly scan our infrastructure using industry-standard tools and fix issues promptly."
- Evidence-backed: "External network scans (OpenVAS and Nmap) run weekly against all production IPs and domains. Our customer-facing web app receives a weekly authenticated DAST scan. Critical findings are remediated within 7 days and high within 30. A summary report from the most recent scan is available under NDA."
Every phrase in the second answer can be backed up with a report, a schedule, or a ticket. Only commit to timeframes you actually meet; a reviewer who later sees a 90-day-old critical finding will remember the 7-day promise.

Build the evidence pack once, then keep it fresh
The fastest way to handle questionnaires is to stop treating each one as a new project. Assemble a small evidence pack that answers the five themes above, store it somewhere your sales and engineering people can both reach, and refresh it on a schedule. HECVAT 4 itself is designed to be completed once per year and shared with any institution, which is a good mindset for every questionnaire: answer once, reuse often, update when reality changes.
A scope statement
Write one short paragraph that lists what you scan: production IP ranges, public domains and subdomains, the main web application, any public API, and (if you scan it) the internal network. Then list what is out of scope and why, such as third-party SaaS tools you use but do not host. Reviewers respect a clearly bounded scope far more than a vague "everything." If you are not sure your target list is complete, an external attack surface review is a good first step before you write it down.
The schedule and its history
A schedule proves intent; a history of completed runs proves follow-through. If the CAIQ asks for at least monthly scanning, a weekly or monthly recurring scan with twelve months of completed runs is the cleanest possible answer. In Panoptic Scans, each scan can be set to run daily, weekly, monthly, quarterly, or yearly (see Running Scans), and recurring scans keep a history you can step through to compare what changed between runs, as described in Scan Results and Reports.
A recent, shareable report
Many reviewers will ask for a copy of a recent scan report. Decide in advance what you are comfortable sharing and under what terms. A common approach is to share a summary and severity counts freely and the full report only under NDA. A good report lists each finding with the affected host and port, a severity score, how it was detected, and the recommended fix, so the reviewer sees real work. Our sample report shows that structure using a real OpenVAS scan of a public test host.
Remediation proof
The most persuasive evidence is a finding that appeared, was fixed, and then disappeared on the next run. Keep a short written remediation SLA by severity, and keep enough history to show you meet it. For deciding what goes first when several issues land at once, our guide to vulnerability prioritization covers using severity alongside signals such as the CISA Known Exploited Vulnerabilities catalog.
An exceptions record
Every real program has findings that will not be fixed: a false positive, a test endpoint behind other controls, a risk the business has consciously accepted. Questionnaires often ask how these are handled. The answer reviewers want is that each exception is recorded with a reason, a decision maker, and a review date. Our article on verifying and recording false positives walks through how to do this without letting the exceptions list become a dumping ground.

Answering the awkward questions honestly
Some questions are uncomfortable for small teams. Honest, specific answers almost always go down better than evasive ones.
"Do you have any open critical or high vulnerabilities?"
Answer with the current state and your process, not just "no." For example: "As of our most recent weekly scan, there are no open critical findings. Any critical finding is triaged within one business day and remediated within our 7-day SLA." If something is open, say so, and say what is being done. Reviewers know software has bugs; what worries them is a vendor unaware of their own.
"Do you perform penetration testing?"
Do not present a vulnerability scan as a penetration test. They are different activities, and CAIQ lists them as separate TVM controls. If you have not had a pen test yet, say that you run continuous automated scanning and state when you plan an independent test. Our comparison of vulnerability scanning and penetration testing explains which one a reviewer expects at which stage of growth.
"Do you scan your internal network?"
For many cloud-native SaaS companies, the honest answer is that there is very little internal network in the traditional sense. Say what you have: cloud workloads, private subnets, office devices, or none of these. If you do run private infrastructure, an internal scan from inside that network is the evidence that answers this. Panoptic Scans offers this on the Pro plan through an OpenVAS scanner you run on your own VM, described in the internal network scanning docs.
"Do you check third-party and open-source libraries?"
This maps to the CAIQ control on external library vulnerabilities, and it is a different discipline from network and web scanning. External scanners can spot some outdated components from the outside, but they do not read your dependency manifests. Answer this one with whatever software composition or dependency alerting your code platform provides, and do not stretch your scanning evidence to cover it.
How MSPs can use the same approach for clients
MSPs see questionnaires from two directions. Their own customers send them, and their clients often ask for help answering the ones they receive from their own customers. A repeatable evidence pack per client turns that second request from a scramble into a billable service.
In practice this means one scope statement and one recurring scan set per client, results delivered to the right contact, and a short quarterly summary the client can attach to any questionnaire. On the Premium and Pro plans, Panoptic Scans can put your logo on PDF reports and add an AI-generated attack narrative that explains in plain language how findings could be chained together, which is useful for non-technical client stakeholders. The MSP vulnerability scanning page covers how providers organize many clients in one account and push results into their own tools through webhooks and the API.
Where Panoptic Scans fits, and where it does not
Panoptic Scans is a hosted service that runs OpenVAS, ZAP, Nmap, and Nuclei, so the scanning, scheduling, and reporting parts of the evidence pack can come from one place. Findings land in a single vulnerabilities view where you can filter by severity, status, and scanner, record a finding as an accepted risk or a false positive, and export to CSV. Findings that are no longer detected on later scans resolve automatically and reactivate if they return, which gives you a natural record of remediation over time (see Vulnerability Management). If your customers or auditors already look at your Vanta account, the official Vanta integration uploads scan evidence there automatically.
It is just as important to be clear about the limits. Panoptic Scans does not fill in questionnaires for you, write your vulnerability management policy, or replace an independent penetration test. It does not analyze your source code dependencies. And scan results only help a questionnaire if your target list actually reflects what you run in production. The tool produces the evidence; your team still owns the scope, the decisions, and the answers.
Turning questionnaires into a sales asset
The first questionnaire is painful because you build every answer from scratch. By the fifth, a team with a current evidence pack can often turn around the vulnerability section in an afternoon. Buyers notice that speed.
If you already run SOC 2, much of this overlaps with the scanning evidence your auditor expects, covered in our guide to SOC 2 vulnerability scanning. If you are starting from zero, begin small: put your production targets on a recurring weekly or monthly schedule, keep the reports, and write down how you handle what they find. You can review plans on the pricing page, and the trial lets you run a first scan against one target before you commit. Then, when the next spreadsheet arrives, you will be answering with evidence instead of adjectives.