← All resources
Business

Antivirus, VPN, password managers, MFA: what actually stops phishing?

“We’ve already got security tools” is the most common reason to skip phishing protection. So let’s be honest about what each tool in the typical stack really does against a phishing attack.

Business · 6 min read · By SafeToOpen Research · June 2026

Walk into most organizations and you’ll find a familiar stack: antivirus, a VPN, a password manager, MFA, a spam filter, and an annual training module. Each is valuable. None of them was built to stop the specific thing a phishing page does — persuade a human to hand over a credential or a token, in real time, on a page that looks legitimate. Here’s the honest breakdown.

Antivirus & EDR

Excellent at catching malicious files and processes on a device. But a credential-phishing page drops no file — the user simply types a password into a web form. Endpoint tools largely can’t see what happens inside the browser tab, which is exactly where modern phishing lives.

VPN

A VPN encrypts your traffic and hides your IP. It does nothing about a phishing site you choose to visit — the connection to the attacker is simply encrypted. “We use a VPN” is one of the most common phishing misconceptions.

Password manager

Genuinely helpful: a good manager won’t auto-fill credentials on a look-alike domain, which quietly stops some attacks. But users can still copy a password in manually, and it offers nothing against QR-code lures, voice scams, or malicious attachments.

MFA

Essential — it blocks more than 99.2% of account-compromise attacks. [1] But attackers stopped beating MFA and started going around it: adversary-in-the-middle kits relay your real login and steal the session token issued afterward, which bypasses MFA, SSO and Conditional Access entirely. [2][3]

Email gateway / spam filter

Catches known and bulk threats well. It struggles with never-before-seen pages, links that turn malicious after delivery, QR codes, browser-in-the-browser pop-ups, and anything that arrives outside email.

Awareness training

The one defense aimed at the human — and the evidence shows its effect is small and fades within months, because the human can always be fooled in the 21 seconds it takes to click. [4]

16%
Phishing is the single most common way breaches begin — the gap every other tool leaves open. IBM, 2025. [5]

The gap they all share

Line them up and the pattern is clear: each tool guards a different layer — the file, the network, the password vault, the inbox, the human — but none of them judges the actual page at the moment of the click. That’s the seam phishing is designed to slip through.

The missing layer

SafeToOpen analyses the page and the email in real time, at the point of click — catching the zero-day phishing site the rest of your stack can’t see. It complements MFA and the tools you already run.

See how it works →

The takeaway

Keep every one of these — they’re all worth having. But owning them isn’t the same as being protected against phishing. The honest question isn’t “do we have security tools?” It’s “what judges the fake page itself, the instant a person is about to trust it?”

See what real-time protection adds

SafeToOpen fills the gap your existing tools leave — judging the page itself, free to start.

See plans →

Sources

  1. Microsoft Learn (Entra) — MFA blocks >99.2% of account-compromise attacks learn.microsoft.com
  2. WorkOS — adversary-in-the-middle & session-token theft workos.com
  3. Obsidian Security — token theft & OAuth abuse bypass MFA/SSO www.obsidiansecurity.com
  4. Verizon 2025 DBIR — median time-to-click, via Stingrai www.stingrai.io
  5. IBM Cost of a Data Breach 2025 — phishing the leading initial vector ($4.8M avg), via Bluefin www.bluefin.com

External statistics are attributed to their original publishers and were accurate at the time of writing. Figures from industry reports vary by methodology and period; we link to primary sources so you can verify them.

test yourself

Could you spot the fake?

Put this into practice: 12 real-world scams and genuine messages, two minutes, no sign-up.