Trust, AI & The Future of Security: Key Insights from Microsoft Secure 2025

At Microsoft Secure 2025, the cyber security landscape took a bold step forward. In a dynamic keynote led by Vasu Jakkal (CVP, Microsoft Security), Charlie Bell (EVP of Microsoft Security) and other senior leaders, Microsoft introduced powerful innovations aimed at securing the future of AI.

From Security Copilot Agents that autonomously triage incidents, to real-time DLP controls for ChatGPT and Gemini, and Defender for Cloud’s AI posture management across Azure, AWS and Google—this wasn’t just a roadmap. It was a call to action.

At CyberOne, we’ve already turned these innovations into real-world defence strategies for our clients. Here’s what it all means—and how to prepare to fight AI-powered threats with AI-powered defences.

Highlights from Microsoft Secure 2025

The Threat Has Evolved. So Must We.

• Microsoft now tracks 1,500+ threat actors and sees 84 trillion signals daily.
• AI is used to write more convincing phishing emails, generate deepfakes, and manipulate identities in real-time.
• New actor techniques include wallet abuse, ASCII smuggling, prompt injection, and custom malware that adapts as it moves.

CyberOne in Action: Our Managed eXtended Detection & Response (MXDR) services, powered by Defender XDR and Sentinel, help you detect and respond to these threats 24×7, with full visibility across endpoints, cloud, identities and data.

Security Copilot Agents: AI That Works With You

Microsoft unveiled Security Copilot Agents—AI-powered assistants that learn, adapt, and operate autonomously to speed up security response while remaining under human control.

Top Use Cases:

• Phishing Triage Agent: Automates review of suspicious emails.
• DLP Alert Agent: Classifies and prioritizes data loss risks based on content and user behavior.
• Vulnerability Remediation Agent: Suggests and applies security patches using Intune and Defender data.
• Conditional Access Agent: Dynamically updates Entra access policies based on new users or threats.

Defending the AI Layer: Shadow AI, Model Abuse & Regulatory Compliance

AI adoption has exploded—and so has risk. Microsoft has responded with new tools within Microsoft Purview, Microsoft Entra, and Microsoft Defender products to manage AI threats across platforms.

• Purview for Edge: Blocks sensitive info from being typed or pasted into ChatGPT, Gemini, and others.
• Entra AI Web Filter: Restrict AI app access by role and risk level.
• AI Security Posture Management: Defender for Cloud now supports OpenAI, AWS, Google Vertex, Azure Foundry (Mistral, DeepSeek, etc.)
• Runtime Protections: Defender XDR now detects prompt injection, credential theft in model responses, and wallet abuse.

CyberOne: Configuring and managing these protections for clients across financial services, public sector, healthcare and high-risk industries.

Accelerated Post-Breach Response with Data Security Investigations

New in Microsoft Purview, Data Security Investigations leverages generative AI to analyse thousands of compromised documents in minutes, mapping risk, intent, and severity for rapid remediation.

Missed Microsoft Secure 2025?

Watch the Microsoft Secure 2025 digital event now and see AI-powered security in action.

Need Help Securing Your AI Strategy?

Speak to the CyberOne security experts and start building a resilient defence strategy—Contact Us.

Don’t Miss Our Upcoming AI & Human Intelligence Webinar

AI & Human Intelligence: The Best Defence Against Cyber Threats:

Cybercriminals Are Using AI—It’s Time to Fight Back Smarter

Join CyberOne’s Microsoft Security leaders as we explore how attackers leverage AI—and how you can respond with AI-driven tools and expert human oversight.

What You’ll Learn

• How hackers use AI for deepfakes, spear-phishing and adaptive malware
• The true cost of AI-driven breaches—and how to avoid them
• How to combine Microsoft Security with AI threat intelligence
• Real-world case studies from our frontline defence teams

Event Details

AI & Human Intelligence: The Best Defence Against Cyber Threats:
📅 Thursday, 1 May 2025
🕙 10:00 – 11:00 BST
🌐 Online Event
🎙️ Speakers:

• Luke Elston, Microsoft Practice Lead, CyberOne
• Lewis Pack, Head of Cyber Threat Defence, CyberOne

REGISTER NOW and learn how to turn AI into your strongest cyber defence.