1. What an endpoint is and why it is the most frequent target
An endpoint is any device that connects to the corporate network and can run programs: desktop PCs, laptops, servers, phones and tablets. They are where a user opens a document, clicks a link or uses an application, so they are also where an attack takes shape. Malware almost always enters through a device, then moves towards shared data and systems. For this reason protecting endpoints is not a separate technical matter, but the first line of defence for the whole organisation.
The point to clarify straight away is that no single protection is enough. A good antivirus does its job against known threats, but it does not cover an outdated device, a weak password or a user who installs a program from an unknown source. Endpoint security is a set of controls that reinforce each other: prevention, detection and response. The work begins with an inventory: knowing how many devices you have, who uses them, what data they hold and whether they leave the office.
2. Basic protections: updates, encryption and accounts
The foundations are four: operating systems and applications kept up to date, encrypted disks, accounts with limited privileges and multi-factor authentication for the most sensitive access. Updates close the vulnerabilities that attacks exploit most often, so a device left behind is an open gap. Disk encryption protects data when a laptop is lost or stolen, but it does not stop malware that is already running. Accounts without administrative privileges limit the damage of a user's mistake. Each point requires a written decision.
The part most often forgotten is the update cycle. Ask your supplier which programs are covered, how often security fixes are applied and how it is verified that they are really installed. A report that says 'up to date' without a list proves nothing. The topic is covered in more depth in the guide on automatic or scheduled updates, which compares the two policies and helps you choose the one that suits your working rhythm.
3. Next-generation antivirus and EDR: what changes
Traditional antivirus recognises files already catalogued as malicious. More recent solutions add behavioural analysis, meaning they watch what a program does: whether it encrypts many files in a short time, whether it changes system settings or tries to switch off other controls. EDR (Endpoint Detection and Response) goes further: it records device activity, allows the device to be isolated from the network and reconstructs the sequence of an attack. Not all products offer the same functions, and commercial labels are no substitute for verification.
The practical difference concerns what happens afterwards. An antivirus blocks; an EDR lets you understand what happened and contain the incident. For a small organisation the useful question is not which technology is best in the abstract, but who watches the alerts and how quickly. An alarm nobody reads is worth no more than no alarm at all. Ask your supplier who monitors the events, which actions it can take without your consent and how it notifies you.
4. Application control and removable devices
Many infections begin with a program that a user downloads or runs without anyone having approved it. Application control, also known as allowlisting, allows only authorised programs to run and blocks everything else. It is a powerful measure, but it must be introduced gradually, because specialist tools used by some departments can be blocked by mistake. Start with a group of devices with uniform use, observe what would be blocked and only then apply the rule.
The same principle applies to USB sticks and external disks. Blocking them altogether is often impractical, but allowing any medium is a risk. A middle path is to define which devices are permitted, who may use them and what data may be stored on them. Managing removable media requires written rules. The dedicated guides on allowlisting and USB media describe the operational steps and are worth reading after this overview.
5. Email and phishing: the entry point almost everyone uses
Email remains the most used channel for bringing an attack into a business. A mail filter reduces the volume of malicious messages, but no filter is perfect, and the message that gets through is often well written, with a credible sender and an urgent subject. Protection must therefore combine technical controls on email, such as link and attachment analysis, with the habit of checking before acting, especially for payments and changes to bank details.
The second element is a clear procedure for reporting. An employee who suspects a message must know whom to contact, through which channel and what not to do, for example clicking or forwarding it to colleagues. Without this procedure, reports arrive late or not at all. The dedicated guide on reporting phishing explains how to build it, while the comparison between a traditional gateway and API-based protection concerns the architecture of the filter.
6. Device management: company PCs, personal laptops and mobiles
A protected endpoint is also a managed endpoint. A centralised management system, often called MDM (Mobile Device Management) for phones and similar tools for PCs, lets you apply configurations, install updates, encrypt disks and wipe data if a device is lost. Without management, every device becomes a configuration of its own, and nobody knows for certain what state it is in.
The most delicate issue is personal devices used for work. The BYOD model (Bring Your Own Device) reduces costs but shifts the line between business and private data. Models with company-owned devices that allow personal use, or with devices used exclusively for work, offer more control. The choice depends on the data handled and the level of trust with staff, and must be decided before buying tools, not after.
7. Monitoring, logs and who responds when something goes wrong
Protection only works if someone watches the signals. The main events to collect are successful and failed sign-ins, software installations, changes to security configurations, devices that stop communicating and alerts from the defence system. Keeping them for a defined period allows an incident to be reconstructed. Ask your supplier where these logs are stored, who can read them and for how long they remain available.
The second element is accountability. You need a name for each stage: who receives the alert, who decides to isolate a device, who informs management and who assesses whether an incident must be notified. The NIS2 directive requires, for those within its scope, a first notification within 24 hours of becoming aware of a significant incident. GDPR requires a personal data breach to be notified to the authority within 72 hours, where the conditions it sets out are met. These deadlines should be known in advance, not during the emergency.
8. Common mistakes in endpoint management
The first mistake is believing that installing a tool is enough. A tool without configuration, without people who read its alerts and without a response procedure produces a false sense of security. The second is treating all devices in the same way: the laptop of someone who travels has different needs from a PC in the warehouse. The third is forgetting servers, which often receive less attention than PCs but hold the most important data. A compromised or stopped server often blocks the entire organisation, not just the person using it.
Another frequent mistake is accepting exceptions without an expiry date. An application excluded from controls for a temporary need stays excluded for years. It is better to record every exception, with the reason, the owner and a review date. Finally, many organisations never verify their controls: a periodic test, even a simple one, shows straight away whether a protection really works or is only configured on paper. The result should be documented and compared with what the supplier has declared, because the two do not always match.
9. How to start
This week, make a minimal inventory: list the devices in use, who uses them, which operating systems and which data they hold, and mark those that leave the office. No special software is needed to start; a shared sheet kept up to date by the IT manager or supplier is enough. With this list you can assess the basic protections without starting blindly. Even a partial list is already useful, as long as it is kept up to date.
Then choose an order. For most organisations it is best to start with updates and encryption, define the procedure for reporting suspicious emails, and only afterwards introduce tighter controls such as allowlisting. Before going deeper, read the hardening checklist for PCs and laptops, which offers a list of concrete checks, and then the guide on updates, which helps you choose a patching policy. Proceed step by step and verify each one before moving to the next, so that every intervention is really closed. Record each step with a date and an owner so you can review it.
Frequently Asked Questions
Is a good antivirus enough to protect company PCs?
No. An antivirus is a necessary but not sufficient control. It does not cover outdated devices, weak configurations or user mistakes, such as installing unapproved software or entering credentials on a fake page. It should be combined with regular updates, disk encryption, accounts with limited privileges and a procedure for reporting suspicions. Ask your supplier which controls it covers beyond the detection of malicious files.
What is the difference between antivirus and EDR?
An antivirus recognises and blocks known threats or suspicious behaviour on a single device. EDR records device activity, helps reconstruct an attack and allows a machine to be isolated from the network. The value of EDR depends heavily on who analyses the alerts and how quickly they can act. If you have no dedicated in-house staff, check with your supplier whether monitoring is included and what it can do on its own.
Does disk encryption also protect against viruses?
No. Encryption protects data when the device is switched off or lost, because without the key the content cannot be read. Malicious software running on a system that is already started and unlocked can access files with the permissions of the user who ran it. For this reason encryption should be seen as protection against loss and theft, and combined with other controls. Find out more in the dedicated guide.
Which devices should be managed centrally?
In principle, all those that access business data: PCs, laptops, servers and phones. Management matters most for devices used outside the office and for those holding sensitive data. There is no universal threshold: it depends on the type of data, the number of people involved and the risk the organisation accepts. The first step is to know which devices exist, including personal ones used for work.
What must I report in the event of an incident?
It depends on the type of incident and on your organisation. For a personal data breach, GDPR requires notification to the authority within 72 hours of becoming aware of it, where the conditions are met. If you fall within the scope of the NIS2 directive, the first report for significant incidents must be made within 24 hours. Check with a consultant or the competent authority which obligations apply to you, and define in advance who decides.
Where do I start if I have little time and a small budget?
Start with an inventory of devices and data, then apply automatic or scheduled updates, encryption on laptops and accounts without administrative privileges for anyone who does not need them. Define a simple procedure for reporting suspicious emails. These measures do not require expensive tools and reduce many common risks. Only after that should you consider more advanced controls, with your supplier's help.