AzureFeeds

Email security presents a complex challenge for individuals and organizations alike. Over the years, attackers have evolved from simple spam campaigns to sophisticated threats including ransomware, identity theft schemes, and carefully crafted phishing scams. Now, malicious actors are armed with Generative AI and are advancing at an alarming pace.

In response, Microsoft Defender for Office 365 has dedicated extensive research and development efforts to making email security smarter, more flexible, and more proactive. This dedication led to the introduction of specialized language intelligence to fight Business Email Compromise (BEC) attacks, announced last year at Ignite 2024 (Microsoft Ignite: Redefining email security with LLMs to tackle a new era of social engineering | Microsoft Community Hub). With that announcement, we offered a significant leap in analyzing suspicious messages using advanced natural language processing, enabling organizations to better detect subtle manipulative emails designed to lure unsuspecting users into revealing confidential data or transferring funds. 

The threats, however, have not stopped with BEC. Phishing attacks are constantly evolving, leveraging new tactics and forms. As part of Defender for Office’s mission to stay one step ahead of these threats, we’re taking the same robust Language AI approach we used for BEC analysis and applying it to a broader spectrum of phishing attacks.  

Today we’re excited to announce Microsoft Defender for Office 365’s new Language AI for Phish model. This model progressively learns from thousands of real-world phishing attempts and analyzes all messages classified as phish. Furthermore, it incorporates advanced Machine Learning and Natural Language Processing (NLP) techniques to read, process, and understand email content the way a human analyst might, yet in a fraction of the time and at an immense scale. Our model has been operational since April 2025, achieving over 99.99% accuracy and blocking 1 million phishing emails daily. 

By advancing our language AI and rigorously training it on phishing email threats, we are further strengthening the comprehensive protections established by our BEC-focused innovations. These enhanced capabilities create an integrated security framework designed to proactively address evolving risks and accelerate response times to emerging threats. Through Microsoft Defender for Office 365’s commitment to continuous improvement, this expanded approach empowers organizations and individuals to maintain a strong security posture in the face of ever-changing cyber challenges. 

Learn More:

To learn more about Microsoft Defender for Office 365’s Language AI capabilities, please read more here or visit our website.