Skip to content
An open dune horizon at dawn. Wide field, breath of distance.
Phishing Defense Stack

The phishing defense stack.

A working defense for AI-generated phishing, BEC, fresh-domain impersonation, and the Microsoft 365 vector.

Phishing in 2026 looks structurally different than phishing in 2016. AI tools generate prose indistinguishable from a real first email at near-zero cost. Per Keepnet Labs, AI-generated phishing is 24 percent more effective than human-written. Per StrongestLayer, click-through has gone from 12 percent to 54 percent. Content classifiers have not kept pace, because the question they are asking (does this look like phishing) has gotten harder to answer from content alone.

A working defense in 2026 has three layers. Authentication on the sender side (SPF, DKIM, DMARC). Reputation-aware content filtering at the provider (Gmail, Outlook, Microsoft Defender). A known-or-pay gate at the inbox door that imposes structural friction on bulk attempts and fresh-domain impersonations. Each layer catches what the others miss; none replaces the others.

The field notes below cover the structural shift in phishing (anatomy, mechanics, why content classification is losing) and the controls that hold across attack classes. Read in order for an end-to-end view, or jump to the note that fits the threat shape you are facing today.

One plan. One price.

Keep your existing Gmail or Outlook. Cancel anytime.

$1.65
per month
Start protecting

Annual on Lightning includes one bonus month. See full pricing.