White paper · 24 September 2026
What automation finds and what a security specialist must decide
Key points
- Automated tools are good at breadth: they find candidate issues quickly and repeatably.
- They are weaker at judgement: whether an issue is real, reachable and important in your context.
- A report is only as good as the validation behind it. Unvalidated findings waste engineering time.
- Say in the report which findings were confirmed by a person, and how.
The division of labour
Penetration testing as a service combines two things that are often confused. Automation maps what is exposed and probes it. A person reads the output and decides what it means. Treating the first as the whole service produces long lists that nobody trusts. Treating the second as the whole service is slow and expensive for what machines do well.
What automation does well
- Discovering hosts, services, certificates and exposed endpoints, repeatedly and at low cost.
- Checking for known weaknesses and misconfigurations across many systems.
- Detecting change: what is new since the last run.
What needs a person
- Confirmation. Is the issue real, or a false positive? Can it be reached from where an attacker would stand?
- Context. A moderate flaw on a system holding personal data may matter more than a severe flaw on an isolated test server.
- Chaining. Several small weaknesses can combine into one serious path. Tools list them separately.
- Explanation. A finding the team cannot reproduce, understand or fix has not helped them.
What a good report states
- Scope: what was tested and what was not.
- For each finding, whether it was confirmed by a specialist, and how it was reproduced.
- A priority with its reasons, so the team can agree or challenge it.
- A mapping to the standards your auditors use, such as ISO/IEC 27001, NIS2 and GDPR, as a reference and not as a certification.
- What to retest after fixes.
Questions to ask a supplier
Who checks the findings, and are they named in the process? How many of the findings in a typical report are confirmed? Can you see how a finding was reproduced? How is a retest handled?
Limits
No test proves that a system is secure. It shows what was found, within a scope, at a point in time. Continuous testing narrows the gap between tests, and human validation keeps the findings credible.
General guidance. It does not describe any individual client engagement.