Updated November 2026.
Most penetration testing companies will tell you they can test anything. The useful question is narrower: can this team find the paths an attacker would actually use in your environment, and can they do it without causing the outage you hired them to prevent? For an office network the second half barely matters. For a plant, a pipeline or a utility it is the whole decision.
This guide covers what to check before you sign: the testing approach, the selection criteria that separate a real penetration test from a relabelled scan, the main types of provider and where each one fits, and the extra questions to ask when the scope includes operational technology (OT) and industrial control systems (ICS).
What to Look For in a Penetration Testing Company: The Short List
- Manual testing, not a scan with a new cover page. Ask what the testers do after the tools finish.
- An approach that matches your question. “Can someone get in?” and “How far can someone get once inside?” are different tests.
- A written scope and rules of engagement before anyone touches the network.
- Experience in your kind of environment, shown by the systems and sites the testers have actually worked on.
- Communication during the test, not a surprise report at the end.
- Help after the report if your team needs it to fix what was found.
Understanding the Approach Taken by Penetration Testing Companies
Penetration testing companies generally work in one of three ways, depending on how much the testers know about your environment before they start. The right one depends on what you need to learn.
Opaque Box
Also known as: black-box or closed-box testing.
What it is: the testers start with no internal knowledge, usually from outside the network, the way an outside attacker would.
Strength: the fastest to run and a fair test of what is exposed from the outside.
Limit: if the testers cannot get past the perimeter in the time allowed, nothing inside gets tested. A clean result can mean the inside was never reached, which is a false sense of security rather than proof.
Semi-Opaque Box
Also known as: grey-box or assumed-breach testing.
What it is: the testers get some background and low-level credentials, as if an attacker already has a foothold.
Strength: skips reconnaissance and spends the engagement on the question that matters most: how far can someone get once they are in? It also tests the security inside the hardened perimeter.
Limit: it can cost more than an opaque-box test.
Transparent Box
Also known as: white-box or open-box testing.
What it is: the testers get full knowledge of the network, its configuration and its documentation.
Strength: the most thorough review, and a good model of an insider such as a disgruntled employee or a contractor with broad access.
Limit: it says less about how a real outside attack would unfold, requires sharing more information, and usually takes the longest.
Which Approach Is Best?
For most organizations, Red Trident finds that a semi-opaque, assumed-breach test delivers the most value. It answers “If a malicious actor got into the environment, how far could they get and what could they do?” rather than only “Can a malicious actor get in?”
Time is the other reason. A standard Red Trident penetration test with active testing runs over roughly one 40-hour work week. Real attackers have no such limit; they can spend weeks learning an environment. Giving the testers some knowledge up front closes part of that gap, so the week is spent finding paths rather than rediscovering the network map.
Selection Criteria for Penetration Testing Companies
Use these criteria in order. A provider that falls short on the first two is not offering a penetration test, whatever the proposal calls it.
1. Manual testing beyond automated scanning
Some firms sell “penetration testing” that is an automated vulnerability scan with a report. Scans are useful, but they miss what needs a person to chain together: a weak credential that leads to a share that leads to a jump host. They also produce false positives that leave you unsure where to start. Ask what the testers do by hand and request a redacted sample report. If the price looks too good to be true, it usually is.
2. A scope and rules of engagement you agreed to
A good provider spends time understanding your business, your systems and your specific concerns before quoting, so no two tests are the same. The result should be a written scope and a rules-of-engagement document covering what is in and out of bounds, testing windows, who to call if something looks wrong, and how sensitive systems are handled.
3. Experience in your environment
Certifications are valuable: they confirm a tester understands the concepts and the method, and they are worth asking about. They do not by themselves show that someone has worked safely on systems like yours. Ask which kinds of environments and devices the individual testers assigned to you have tested, not what the company has done in general.
4. Communication during the engagement
You should know what the team is doing and what it has found while the test runs. Serious findings should reach you the day they are confirmed, not weeks later in a PDF.
5. A report people can act on
Look for findings ranked by consequence to your business, the evidence behind each one, and remediation steps your team can carry out. A long list sorted only by a generic severity score is harder to use.
6. Support after the report
Decide whether you want a one-time test or a partner that can help fix what was found. If you have your own team, a good provider will step back, support the parts you want help with, or train your staff.
Types of Penetration Testing Companies Compared
Penetration testing providers fall into a handful of types. Each has a place; the mistake is hiring one type for a job that needs another.
| Provider type | Strengths | Watch for | Best fit |
|---|---|---|---|
| Automated or “pen-test-as-a-service” platforms | Fast, inexpensive, repeatable | Mostly scanning; little manual chaining of findings | Frequent checks of internet-facing IT systems between full tests |
| Large IT security consultancies | Broad bench, compliance-driven reporting, global reach | Testers vary by engagement; OT work may be subcontracted or handled by IT testers | Enterprise IT, applications and compliance-driven programs |
| Boutique IT penetration testing firms | Deep manual skill, senior testers on the work | May not know control systems or plant constraints | Web applications, cloud and corporate networks |
| OT security product vendors with services arms | Strong knowledge of their own monitoring platform | Findings can lean toward the vendor’s product | Organizations already standardized on that platform |
| OT/ICS specialist firms, such as Red Trident | Control-system experience, safety-first methods on live production | Smaller teams; check availability and the experience of the assigned testers | Plants, utilities, pipelines and other operational environments |
For a deeper comparison of providers for industrial environments, see Selecting ICS/OT Pen Testing Providers for Industrial Assets.
Penetration Testing Companies That Specialize in OT and ICS
If your scope includes operational technology, the selection criteria above still apply, and four more matter as much.
Consequences, not just access
An OT test should show the paths an attacker could take toward the control network and whether they could cause OT-specific consequences: loss of view, loss of control, loss of confidence in what operators see, inhibited response functions, or impaired process control. “We got domain admin” is not the end of an OT test.
How the team tests a running system
Ask for the method in order. A sound approach starts with passive observation, which shows what is actually happening on the network: which devices talk to which, over which protocols, and which flows cross zone boundaries. It then moves to active testing, which shows what is exposed and reachable: open services, exact versions, configuration weaknesses and the paths between them. Interviews and walk-downs find the informal connections and dormant devices that neither method sees. Active testing on a production control network is safe when it is done by engineers with hands-on control-system experience who understand the potential impacts and plan around them. Vet the team’s experience; do not settle for a scope that avoids active work.
On-site work for live production
Red Trident prefers to test on site, especially for live production environments. Being at the plant means coordinating with operators and shift supervisors in real time and stopping immediately if something looks wrong.
Testing windows and availability
Availability comes first in OT. The provider should agree on testing windows with your operations team, know which systems must not be touched during production, and say exactly how it will avoid affecting the process. Not every penetration testing company will respect those constraints, or even be aware of them.
Services Beyond Penetration Testing
A penetration test shows where you are exposed; it does not fix anything. If your team needs help afterward, ask whether the provider also offers vulnerability assessments, asset discovery and remediation. Red Trident does, and is equally happy to step back and let your team handle the fixes, or train your staff to do them.
Penetration Testing Companies: Frequently Asked Questions
What is the difference between a penetration test and a vulnerability scan?
A vulnerability scan is automated and lists known weaknesses on the systems it can reach. A penetration test uses people to exploit and chain those weaknesses to show what an attacker could actually achieve. Scans are a useful part of a program; they are not a substitute for a test.
How long does a penetration test take?
It depends on scope. A standard Red Trident penetration test with active testing runs over roughly one 40-hour work week, plus scoping beforehand and reporting afterward.
Which approach should we choose: opaque, semi-opaque or transparent box?
For most organizations, a semi-opaque, assumed-breach test gives the most useful answer for the time spent. Choose opaque-box if your main question is external exposure, and transparent-box if you want the most thorough review or are modelling an insider.
Can our internal security team do the penetration test?
Sometimes, if they have the skills, tools and time. Internal teams often find it hard to take an outside attacker’s view of systems they built and maintain, and the specialized tooling and training are expensive to keep current for occasional use. An external test also gives auditors and leadership an independent view.
Is it safe to penetration test an operational control network?
Yes, when it is done by engineers with hands-on control-system experience, following a method that starts with passive observation, agreed testing windows and clear rules of engagement. The risk comes from testers who do not understand the process they are working near.
Do certifications matter when choosing a penetration testing company?
Yes. Certifications confirm that testers understand the concepts and methods. Ask about them alongside the real test of fit: what systems and environments the testers assigned to you have actually worked on.
Choosing Your Penetration Testing Company
Start with the question you need answered, choose the approach that answers it, and hold every provider to the same criteria: manual testing, an agreed scope, relevant experience, communication during the test, an actionable report and support afterward. If your environment includes control systems, add the OT questions above. To talk through what a test of your environment would look like, schedule a call with Red Trident from our ICS/OT penetration testing page.
