Here's a phishing campaign that should worry every security leader more than the big ones that make headlines.
In August, our SOC team caught a campaign that reached 10 client tenants without a spoofed sender, a compromised account, or a single newly registered domain.
The attacker didn't break in anywhere, they just signed up for a free trial on a legitimate, well-established education CRM platform, then used that platform's own notification system to send phishing at scale.
It passed every reputation check with ease.
The sending domain had roughly a decade of clean DNS and WHOIS history. The mail came from real, shared sending infrastructure with a reputation the attacker did nothing to earn; they just rented 10 minutes of it. Even the subject lines were personalized for each company, not sprayed from a single template. This was an aimed attack.
And here's where it gets really uncomfortable: nothing about this was a flaw in Microsoft's tooling, or our tech stack, or the platform that got abused. Every system did exactly what it was built to do, it just wasn't built to answer the right question.
The industry has spent two decades scoring infrastructure. Attackers have started scoring intent instead.
Sender reputation, domain age, and IP history are all measures of the past because they only tell you whether something has behaved well before. They tell you nothing about what it's asking you to do right now.
A platform that's sent legitimate mail for years keeps its clean record on the exact day someone else uses it to phish your employees. The infrastructure isn't lying, it's just answering a different question than the one that matters.
What finally caught this campaign wasn't a filter, it was a person who reported an email that felt off, correlated against nine other reports that individually looked like isolated, low-priority spam tickets. Looked at one inbox at a time, this campaign would never rise above the noise. Looked at across tenants, it's obviously one operation.
That's the part I keep coming back to, and which shows the strength of RADICL. We've built enormously sophisticated detection systems, and the thing that closed this loop was still a human noticing that a message wanted something it shouldn't.
Three things I'd ask every security and business leader to sit with:
The attacker in this story didn't need to be sophisticated. They needed 10 minutes, a card, and a signup form. The barrier to entry for this kind of attack is falling, and it will keep getting lower. The barrier to catching it needs to fall just as fast, and that starts with admitting that "it passed every check" and "it's safe" are two different sentiments.