Abnormal AI

AI made cybercriminals more effective. Even worse, it made more people into cybercriminals. More attacks. More sophistication. Fewer stopped by conventional security. Evan Reiser on why these dimensions multiply into a perfect storm for the bad guys.

Introducing Hyperdrive - a daily build journal from Abnormal's product and security teams. This week: identity proof-of-value, lateral movement via non-human identities, insider threats that pass background checks, and more. First issue is live. Featuring contributions from: Shaun Archer Stephen Harrison Brandon Stacey Subscribe to see how and what we're building.

The Abnormal After Hours party at Infosecurity Europe did not disappoint. 💥 We had so much fun building community with our customers and partners at the Abnormal After Hours party last night — and we're not done yet. Stay tuned for the full Infosec Europe recap dropping Monday.

AI Phishing Coach wins 2026 Cloud Security Innovator of the Year! The Security Awards, run by global cloud computing awards body The Cloud Awards, recognizes landmark achievements in cybersecurity across more than 20 categories. Traditional security awareness training has tried to solve a people problem with rules. AI Phishing Coach takes a different approach. By learning each employee's behavior, we deliver personalized, adaptive coaching that builds real resilience, beyond just compliance scores. The judges saw it the same way, saying: "Abnormal AI stood out as a forward-thinking and highly effective advancement in security training and human-centered cyber defense." Learn more about how AI Phishing Coach is redefining what security awareness looks like: https://lnkd.in/e7uv4rRK

We're on the ground at Infosecurity Europe! As we wrap up day one, we're so excited for what's to come the rest of this week. From hosting meetings at our booth, to driving industry conversations forward with a speaking session from Patricia Titus, and celebrating our customers on Wednesday at our Abnormal After Dark Party. If you're in London, come find us. Book a meeting here: https://lnkd.in/eQRPHjYB

The Adecco Group, the world's largest workforce solutions company, was protecting 82,000+ mailboxes across 62 countries with Microsoft Defender, which was catching just 15–20% of threats. Everything else landed in inboxes. Manual triage covered only 10–15 executives. The broader workforce had no visibility, no feedback, no safety net. The fix was AI-native protection that could accurately detect advanced, socially-engineered attacks and automate response across the organization. After deploying Abnormal's Inbound Email Security and AI Security Mailbox, Adecco transformed how it detects, triages, and responds to threats across three global business units: 🛡️ 205,000+ advanced attacks autonomously remediated 📬 6,900+ user reports automatically triaged across multiple languages 💰 $8.7M in total risk avoidance Alex Gomez, Global Head of IT Security, Risk, and Compliance, calls it a measurable "cultural shift." One that even earned praise from Adecco Group's CEO for being "so quick and accurate." Read the full story → https://lnkd.in/eyXaVD8P

May at Abnormal looked a little like this: strengthening partnerships on the golf course, taking the stage in Sydney, and celebrating our top performers in Hawaii. 🌺 We took Denver and Vegas by storm, cheered in Atlanta, and somewhere in between, a whole lot of work got done. This is what it looks like to build a global company where the hustle shows up in the boardroom, the ballpark, and everywhere in between. Here's a peek into #LifeAtAbnormal this month.

AI detection systems are getting better at catching threats, but most don't show you what they caught or prove your team's input actually shaped the outcome. When analysts report a missed attack, protection should improve. Our Detection 360 now closes that loop. 🗂 Every submission has a full case history—from initial classification through investigation, remediation, and deployed detections. 📬 Each deployed detection shows a running count of messages caught, tied directly to the report that triggered it. 📊 Program leaders get aggregate visibility across all feedback-driven detections: detections deployed, messages caught, trends by submission type, impact by attack type. Detection accuracy and explainability aren't mutually exclusive and Detection 360 is how we deliver on that commitment. Read more: https://lnkd.in/gaBd3QXQ

According to CISA, over 90% of cyberattacks begin with email. And federal agencies are specifically targeted. 🎯 Attackers study how agencies operate: procurement workflows, org structures, inter-agency communication patterns. Then they build attacks around them. Read more from Yejin Jang on why generic email security wasn't built for this: https://lnkd.in/e_eMRh5D

For the fourth consecutive year, Abnormal AI has been named to Redpoint Ventures' InfraRed 100, an annual list recognizing the private companies building foundational infrastructure for the AI era. The companies on this list aren't retrofitting AI onto legacy architectures — they're building from the ground up around it. That's been Abnormal's approach since day one: training our own behavioral foundation models, learning from threat patterns across our customer base, and detecting attacks that no rule or signature could anticipate. Four consecutive recognitions reflects what happens when you make the right architectural bet early and keep compounding on it. https://lnkd.in/gVRDNS6Q