← All posts
Managed ITThreat detection

Real-Time Threat Detection: Keeping Your Business Safe

Some things can be stopped in seconds by software. Others need a person. It helps to know which is which.

Ransomware on a single computer can encrypt files faster than anyone could walk across the office to unplug it. A stolen Microsoft 365 password can be used from the other side of the world within minutes of a phishing page capturing it. Against attacks that move at that speed, a security alert that someone reads tomorrow morning doesn’t help much.

That’s the case for real-time threat detection. The phrase gets used loosely in sales material, though, so it’s worth pinning down what it can and can’t do.

What “real time” actually means

In practice there are two speeds.

  • Machine speed is seconds. Software sees something it’s confident is malicious and acts on its own: stops a process, quarantines a file, cuts a computer off from the network, signs a user out. No person is involved until afterward.
  • Human speed is minutes to hours. Software sees something suspicious but not certain, raises an alert, and a person looks at it, decides what it means and acts.

A good setup uses both. The obvious attacks are stopped automatically. The ambiguous ones go to a person quickly, with enough information to decide. “Real-time detection” that only sends an email to an inbox is the second kind, without the person.

What can be stopped automatically

Modern tools can take several actions on their own when they’re confident.

On a computer

Endpoint detection and response (EDR) software watches what programs do. If one starts encrypting files in bulk, or behaves like a known attack tool, EDR can stop it, quarantine the file and isolate the computer from the rest of the network while still letting the security team connect to investigate. Isolating one laptop is a lot better than losing the file server.

On an account

Identity protection watches sign-ins. If a sign-in looks risky (an unfamiliar country, a known malicious address, a pattern of password guessing), it can block the sign-in, require multi-factor sign-in again, or lock the account and end its active sessions.

In email

Email filtering can pull a message out of every inbox after it’s been delivered, once it’s found to be malicious, including from people who haven’t opened it yet.

At the edge of the network

Firewalls with intrusion prevention block traffic that matches known attacks as it arrives, and web filtering stops staff reaching known malicious sites.

What still needs a person

Automation is right to be cautious. A tool that locks accounts on every unusual sign-in would lock out the salesperson traveling for work every week. So plenty of alerts are left for a person to judge:

  • Is this sign-in from Atlanta the employee at a conference, or someone with her password?
  • Is this new mail forwarding rule something the user set up, or an attacker hiding replies?
  • Was this file really malicious, or a legitimate program behaving oddly?
  • Is this one incident, or part of something bigger?

And once something is confirmed, people do the rest: find out how it got in, check what else was touched, reset what needs resetting, restore from backup and close the gap.

Isolating one laptop is a lot better than losing the file server.

Why speed matters

Ransomware shows up in a lot of breaches. Verizon’s 2026 Data Breach Investigations Report found it was involved in 48% of the breaches it analyzed (Verizon DBIR 2026). The difference between a ransomware incident that costs an afternoon and one that costs weeks is often how far it spread before it was stopped. One isolated laptop gets rebuilt. A whole office, with the server and the backups that were reachable from it, is a different conversation.

Email compromise works on a slower clock but the same principle. An account that’s locked within minutes of a suspicious sign-in hasn’t had time to read a week of email and send a convincing fake invoice.

Questions to ask your IT provider

If someone tells you that you have real-time threat detection, these questions will tell you what you really have:

  1. What happens automatically, without a person, when a computer is infected?
  2. Can you isolate a computer remotely, and how quickly?
  3. What happens when there’s a risky sign-in to Microsoft 365?
  4. Who reads the alerts that aren’t handled automatically, and how soon?
  5. What happens to an alert at 2 a.m. on a Saturday?
  6. When did you last test this, and what did you find?

The fifth question is the one that separates the setups. Get a specific answer, in writing, and make sure the automatic actions cover the hours when nobody is reviewing alerts.

What you can do yourself

Even before any of this, a few things shorten the time to detection:

  • Turn on multi-factor sign-in for every account, so a stolen password alone isn’t enough.
  • Tell staff exactly who to call if something looks wrong, and make it clear they won’t be in trouble for calling.
  • Keep backups somewhere ransomware on your network can’t reach, and test them.
  • Make sure your provider has a phone number for someone at your company who can make decisions.

For how detection fits with the rest of your security, see why network threat detection matters, and for where the technology is heading, the future of automated threat detection.

How we respond

As part of our cybersecurity service for Managed IT customers, every computer runs endpoint detection and response, with alerts reviewed by our team and threats isolated. Identity threat detection and response watches for unusual sign-ins and locks the account down, and email filtering keeps working on messages after they reach the inbox. If something does get through, we help you contain it, lock down affected accounts and devices, work out what happened and restore from backups. Call (386) 252-2292 if you want to know how your current setup would handle that Saturday-night alert.

More on Managed IT

Let's build your foundation.

Tell us where your business is today and we'll help map the next step.