Exposure Validation Finder

Lesson 1 of 8

What exposure validation is

Exposure validation is testing whether the weaknesses you know about, and the controls you rely on, behave the way you think they do when someone actually attacks.

The gap it fills

Most security programmes are good at listing problems. A vulnerability scanner finds thousands of CVEs. A cloud posture tool flags hundreds of risky settings. An identity review finds accounts with more access than they need. Each list is accurate and each is too long to fix in full.

The same programmes are weaker at proving things. Nobody can say with confidence whether the EDR on the finance laptops would stop credential theft today, or whether the medium-rated flaw on a forgotten file server is the first step toward the domain controllers. Those are questions about behaviour, not inventory, and an inventory tool cannot answer them.

Validation answers them by trying. A test runs an attacker technique, or a chain of them, in your environment, and records what happened. The result is evidence: this technique was blocked, this one was logged but nobody was alerted, this chain reached a privileged account in four steps.

Two kinds of evidence

Validation tools produce two broad kinds of evidence, and the rest of this course keeps them apart because buyers often mix them up.

The first is evidence about controls. Did the email gateway filter the malicious attachment? Did the EDR stop the process injection? Did the SIEM rule fire? This is the territory of breach and attack simulation. It tests one technique at a time, from a known starting point, and grades each control.

The second is evidence about paths. Starting from one compromised machine, how far can an attacker get, and through which weaknesses? This is the territory of automated pentesting and attack path analysis. It cares less about which control failed and more about whether the chain completed.

Gartner's category, adversarial exposure validation, covers both. Its first Market Guide for the category appeared on 11 March 2025, and Gartner Peer Insights describes the category as replacing the older breach and attack simulation and automated pentesting categories. In practice, most products are stronger at one kind of evidence than the other, whatever the label says.

What validation is not

It is not scanning. A scanner looks for known weaknesses by checking versions and configurations. A validation tool acts on them, or acts like an attacker would regardless of them.

It is not a compliance audit. An audit checks that a control exists and has a policy. Validation checks that the control does its job against a specific technique on a specific day. Auditors increasingly accept validation output as evidence, but the two answer different questions.

It is not a replacement for skilled people. A human red team still finds things no automated playbook will: business logic flaws, social engineering routes, odd trust relationships. Validation tools are best at the repetitive, broad coverage that people find tedious and expensive to repeat.

Why repetition matters

The most useful property of a validation tool is that it can run the same test again. Environments drift. A firewall rule is loosened for a migration and never restored. An EDR policy is switched to audit mode while a team chases a false positive. A log source stops sending after an upgrade. None of these announce themselves.

A weekly or monthly test run catches drift as a change from the last result. A yearly pentest catches it only if the timing happens to line up. That is why vendors talk about "continuous" validation, and why the finder asks how often you need to test.

Repetition also makes fixes measurable. When a team changes a setting, the same test can confirm the change worked. Without a retest, a closed ticket only proves someone changed something.

Where the evidence goes

Evidence on its own does nothing. The point of validation is to change what gets fixed and in what order. A good programme feeds validation results into three places: the team that tunes controls and detections, the team that prioritises vulnerabilities, and the people who report risk upward.

Each of those teams wants a different view. Detection engineers want the technique, the expected alert and the log evidence. Vulnerability managers want to know which findings sat on a working attack path. Executives want a trend they can compare quarter to quarter. Later lessons cover each of these in turn.

What to take from this lesson

Before you look at any vendor, write down the one question you most need answered. "Do our controls work?" points toward control testing. "Could an attacker get to our crown jewels?" points toward path testing. "Both, and we need to prioritise everything else around the answer" points toward the broader platforms.

That question is the first one the finder asks, for good reason. The next lesson describes the six tool types that answer it in different ways.


Next: The six tool types.