Skip to content

Incybersecurity, simply detecting a suspicious event isn’t the end of the story. You still need to determine what happened, how far-reaching the consequences are, and what actions will allow you to regain control of the situation. This is precisely what EDR is designed for—a technology that combines device activity monitoring with threat analysis and incident response tools. Its importance is easiest to understand when, instead of asking only about blocking a dangerous file, we start asking about the course of the entire attack.

What is EDR, and what devices does it cover?

The acronym EDR stands for “Endpoint Detection and Response,” which refers to the detection of and response to threats on endpoints, such as employee computers or servers. At the heart of this approach is not only the assessment of individual files, but also the collection of information about events, the reconstruction of their connections, and support for the security team’s activities.

A typical EDR system uses a component running on the device—called an agent or sensor—and a central analysis and management environment, often hosted in the cloud. area The available features depend on the specific solution, operating system, and license; even a critical action such as isolating a device has specific technical requirements. For this reason, a description of such an EDR deployment in a given organization should answer the question of which devices are actually protected, rather than simply stating the number of licenses purchased.

EDR – Zero-day

Antivirus and EDR address different needs

The contrast between “antivirus software detects files, while EDR analyzes behavior” is an oversimplification. Modern Microsoft Defender Antivirus also monitors the behavior of processes, files, and services and detects threats that do not require a match to a known signature. What sets EDR apart, however, is its extensive analytical toolkit—including device activity history, correlations between events, incident investigation, and the ability to take remote remedial actions.

Antivirus and EDR can therefore serve complementary functions, and endpoint protection platforms—referred to as EPP—are increasingly combining threat prevention mechanisms with detection and response capabilities. From an organization’s perspective, what matters more than the name of the solution is whether, once a problem is detected, its cause can be traced and its impact mitigated. After all, a simple block notification does not answer the question of whether similar activity has occurred previously on other computers.

How does the EDR system work?

The EDR process can be viewed as a series of interrelated tasks, such as data collection, data analysis, determining the significance of detected events, and response. Each of these stages has a different objective. A large amount of recorded information does not necessarily equate to an accurate assessment of the situation.

From Telemetry to Event Visualization

Telemetry refers to data describing a device’s activity; in EDR solutions, it may include running processes, network connections, user logins, and changes to files and system configuration. Combining this information enables the analysis of the sequence of events and the relationships between processes, rather than just a review of isolated messages.

However, EDR should not be equated with a complete recording of everything that happens on a computer. When analyzing an incident, it is therefore important to distinguish between a lack of evidence and evidence that an event did not occur. A gap in the timeline should prompt an examination of the scope of available data, not an automatic closure of the case.

From Suspicion to Investigation

EDR software can combine detection rules, behavioral analysis, machine learning, and known threat intelligence to flag activity that requires investigation. Related alerts can then be grouped into an incident, making it easier for an analyst to assess the broader course of an attack.

Let’s imagine a script that organizes directories at a set time and sends a report to the company’s service. In this scenario, its operation is part of routine administration. The same set of actions would warrant a different assessment if the script ran outside of a scheduled time, on an account that doesn’t have such responsibilities, and with an unknown destination address. This hypothetical comparison illustrates why the question “What was run?” should be supplemented with the questions “Who, when, where, and for what purpose?”

From Alert to Response

Depending on the product and configuration, EDR allows you, among other things, to stop a suspicious process, place a file in quarantine, collect evidence for analysis, and isolate the device from the network. Isolation does not necessarily mean shutting down the computer. For example, Microsoft Defender for Endpoint limits the device’s network communication while maintaining a connection to the security service.

Automation can speed up operations, but area the permitted operations must be tailored to the importance of the protected system; typically, these will be restrictions on actions that have a significant impact on the functioning of particularly critical resources. In a hypothetical lab, disconnecting a single workstation would mean an interruption in the work of a single team. A similar decision regarding a server serving all labs would require consideration of much broader consequences.

EDR - Data Breach

A Cyberattack on a Design Studio—An Example of How EDR Works

Let’s consider a hypothetical scenario here. A plug-in was installed in a 3D design studio to streamline the export of models. After the plug-in is launched, the computer begins executing an additional script, a new mechanism for automatically launching the program appears, and then attempts are made to send files unrelated to the current project.

In this example, no single observation should replace an analysis of the whole. The add-on may legitimately use scripts, and the application may communicate with an external service. The problem lies in matching actions with their origin and scope—why does the model export tool access another team’s documentation and attempt to ensure it runs after the computer is restarted?

Let's assume that the deployed EDR logs the necessary events and its rules detect this sequence. An analyst could then investigate the relationship between the application, the add-on, and the script, and, once the threat is confirmed, isolate the workstation and search for similar activity on other computers.

This example also illustrates the difference between halting operations and resolving the incident. Even after isolating the workstation, questions remain regardingarea the exposed data, other installations of the add-on, and the reason it was allowed into the environment. The value of EDR here, however, lies in its support for the investigation, not in the promise of an automatic answer to every question.

EDR and Zero Trust – Device Status Matters

The Zero Trust model does not grant trust simply because a user or device is inside the corporate network. It requires access verification that takes into account identity, permissions, and context, including the device's status.

EDR can provide information that is useful for such an assessment. For example, the integration of Defender for Endpoint with Intune and Microsoft Entra allows you to use a device’s risk level to restrict access to the organization’s resources.

This can be illustrated by a hypothetical scenario involving access to a project repository. An employee logs in correctly and has the appropriate role, but his computer is currently under investigation regarding a suspicious process. In such a scenario, the organization may want to temporarily deny access from that device, even though it does not question the employee’s identity itself.

EDR is therefore not synonymous with Zero Trust, but it can provide information to one of the components of a broader access control architecture. This clearly illustrates the difference between observation and decision-making. The detection system flags a risk, while the established policies determine what consequences that risk has for resource access.

EDR - deepfake

Deepfakes and Cryptography – The Limits of Device Security

Deepfake technology can be used to impersonate others and amplify social engineering attacks; therefore, protection also requires independent verification procedures, not just an assessment of the credibility of a voice or image. area EDR, on the other hand, primarily concerns device activity and responding to detected threats, not determining the authenticity of the other party.

Let’s assume that a doctored voicemail prompts an employee to install a purported update for a design tool. In such a scenario, it is important to distinguish between two questions: whether the command was authentic, and whether the program that was launched performs dangerous actions. One should not expect that detecting these actions will replace the command verification procedure. Moreover, one cannot assume that simply carrying out a flawed business decision will be recognized as a cyberattack.

Another aspect of security is explained in our other article,“What Is Cryptography?”, which covers, among other topics, encryption, hash functions, and digital signatures. These mechanisms serve different purposes—encryption protects confidentiality, while a digital signature allows for the verification of data integrity and its association with a specific key, without in itself guaranteeing the authenticity of the signed content.

In the context of a design studio, it would therefore make sense to ask two separate questions: whether the document comes from the expected source and has not been altered, and what the application does after opening it. A positive answer to the first question does not automatically answer the second. This distinction helps us think of cryptography and EDR as complementary security measures, rather than substitutes.

EDR, XDR, SIEM, and MDR—Similar Acronyms, Different Roles

When designing security measures, technologies that incorporate more data sources and services that support the work of specialists also come into play. They do not form a simple ladder where each subsequent solution replaces the previous one. Nevertheless, we can organize them according to the questions they are meant to address:

  • EDR – What Happens on Endpoint Devices? How Can You Investigate Suspicious Activity and Respond to It?
  • XDR (Extended Detection and Response)—how can you correlate signals from devices with information from other areas, such as email, the network, or cloud environments, and coordinate a response?
  • SIEM, Security Information and Event Management—how can we collect and analyze security data from multiple systems, including to provide broader context for EDR analysis?
  • MDR (Managed Detection and Response)—who will provide the service of monitoring, detecting, and responding to threats using the appropriate tools and a team of analysts?

Distinguishing between technology and service has practical significance. When comparing offerings within a given organization, it is worth evaluating the capabilities of the tool separately and the responsibilities of the incident response team separately ( area ). After all, issuing an alert and carrying out a full response are not the same obligation.

Why isn't EDR software alone enough?

In its recommendations based on its handling of an actual incident, CISA emphasized the importance of continuously reviewing alerts, securing affected systems, updating vulnerable software, and practicing response procedures. The organizational takeaway is therefore simple: purchasing a tool should be the start of building response capabilities, not the end.

When evaluating the implementation, you must first determine area the monitoring parameters—that is, which devices are transmitting data and whether the system will detect if a sensor becomes inactive. Next, it’s important to specify who is responsible for alerts: who reviews reports, prioritizes them, and handles issues outside of standard business hours. 

The limits of automation are also important—that is, the distinction between actions the system can perform on its own and those that require approval. It is also necessary to determine who has access to telemetry data and how long it is needed for analysis. Finally, it is necessary to specify what conditions must be met before the device returns to normal operation.

A useful complement to alert handling is threat hunting—that is, the active search for traces of threats that automated mechanisms have not yet detected or removed. This allows room for the analyst’s hypotheses, rather than limiting their work to responding to a predefined queue of tickets. In the scenario described earlier, such a question might be, for example: Was a similar add-on also launched on workstations that did not trigger an alert?

EDR - a man working at a computer

Skills Needed to Work with EDR

Analyzing incidents in a design lab would require combining several perspectives—application behavior, the operating system, network communication, and the significance of the data being processed. Simply knowing the console interface would not be enough to explain why a particular behavior is dangerous. It is therefore worth treating EDR operation as an application of broader IT knowledge rather than an isolated product-specific skill.

We strive to provide this broader perspective to our students in the “computer science ” program atPJAIT, whose curriculum covers programming, systems architecture, data analysis, artificial intelligence, and cybersecurity. For those interested in further technical education, the postgraduate program in Cybersecurity of IT and Telecommunications Systems is a good option. In collaboration with STM Cyber, we use the HackingDept environment in this program for hands-on training related to attacks and infrastructure protection.

However, incident response also has an organizational dimension. In our hypothetical example, someone would have to decide to suspend operations in the lab, set priorities, and inform the teams about the restrictions. Skills in this area are developed through the postgraduate program in Cybersecurity—Management, Systems, and Product Security, whose curriculum combines technical topics, risk analysis, security management, and DevSecOps, including hands-on analysis of logs and incidents.


From Visibility to a Responsible Response

It’s worth thinking about EDR not in terms of the number of alerts, but in terms of the questions an organization can answer thanks to it. Do we know where the suspicious activity came from? Can we determine its scope, mitigate its effects, and justify the decision to restore the device to operation?

It is precisely this transition from observation to a conscious decision that gives meaning to the entire implementation. Technology is meant to provide information and opportunities for action, while the team must know when and for what purpose to use them. In a well-designed security system, threat detection is not the end goal, but rather the beginning of a structured response.


See other news