SPLX

The AI security landscape is evolving fast - and so is SPLX. 🚀 Our automated red teaming platform has already transformed how enterprises uncover critical vulnerabilities in their AI systems. 🚀 In Q2 alone, we delivered 160% growth and onboarded 5 new Fortune 500 customers. Now, we’re entering the next chapter. Our platform has evolved, and so has our brand identity, reflecting our commitment to securing the entire AI lifecycle. Here's what’s new. 🆕 AI Runtime Protection: Real-time guardrails that act like a firewall for AI apps. Prompt injections, jailbreaks, and sensitive data leaks are stopped the instant they happen. 🆕 Analyze with AI : Turns red team findings into clear, actionable insights so security teams can prioritize and respond fast. With these new capabilities, we are raising the bar for AI security once again. SPLX lets your org move fast and stay ahead with AI adoption - without compromising on safety and security. We’ll be showcasing our platform at Black Hat Vegas 2025. Can’t wait until then? Learn more about SPLX 2.0 👉 https://lnkd.in/dpnDxrmS

Mark Zuckerberg, did you try turning it off and on again? We’ve all had those awkward moments where tech lets us down, but usually not with the whole world watching. That’s exactly what happened when Meta unveiled their AI glasses live on stage this week. When it comes to enterprise AI, failures can damage reputations, leak sensitive data… or, like in this case, give the internet a field day. That’s why the SPLX platform runs deep Q&A tests to catch issues like prompt misalignment, context drift, or instruction failures before launch AND while live in production. And hey, be sure to always double-check that Wi-Fi connection. Thanks Dorian Schultz for the kind offer to help 😉 🎥 Daniel Miessler 🛡️ Joseph Thacker Julie Tsai Chenxi Wang, Ph.D. Michael Sutton Kristian Kamber Ante Gojsalic Edward Amoroso Dennis Xu Karol Lasota Bastien Eymery 🤖

🔥 AI is everywhere, do you know where your company is using AI? Audits, lawsuits, and regulatory pressure are rising and one question keeps catching organizations off guard:  In my latest article, Avoid Lawsuits from Enterprise AI: Why Asset Management Is Your First Line of Defense, I break down: How “AI creep” (vendors, embedded tools, APIs, etc.) exposes you to bias, compliance, and security risks.  What AI asset management looks like in practice—from models to prompts to infrastructure.  Concrete frameworks (AIUC-1, EU AI Act, others) you can apply to classify risk, assign ownership, and shore up governance.  Metrics that matter: visibility, risk reduction, compliance readiness, cost optimization.  📌 Read the full article https://lnkd.in/gBvn9nyR Would love to hear: what’s the biggest blind spot your organization has found when it comes to tracking AI use & risks? Thank you to Barry Hurd for his HR TECH AI COMPLIANCE RISK ASSESSMENT Report used in this article Kristian Kamber Michael Sutton David Endler Stanislav Sirakov Petar Tsachev Lars Godejord Jure Mikuž Manoj Apte Joseph Thacker Sergej Epp Daniel Miessler 🛡️ Miessler John Stewart Ofer Ben-Noon Saša Zdjelar Julie Tsai Luka Kamber Edward Amoroso Dionisio Zumerle Dennis Xu #EnterpriseAI #AICompliance #AISecurity #AIAssetManagement #CISO

Newer ≠ safer. AI capabilities are maturing, but security is often left in the dust. Our CEO and co-founder Kristian Kamber sat down with Michael Vizard at Techstrong TV to discuss what’s going wrong - and why AI red teaming has never been more critical. 👇

🚨 Introducing AI Asset Management – Full Visibility Across Your AI Stack AI is embedding itself into critical business workflows, driving efficiency and unlocking new capabilities. But most enterprises are held back by a blind spot: They don’t know what models, agents, or workflows are running across their stack. AI Asset Management closes that gap. Built on the foundation of our popular open-source project, Agentic Radar, this new enterprise-grade feature delivers full visibility and security for your AI stack. Here’s what teams can do with it: ⚡ Agentic Workflow Discovery Visualize complex agentic workflows across agents, nodes, and tools ⚡ Agent-Level Threat Analysis Detect vulnerabilities with real-time, with benchmarked risk scoring ⚡ Automated AI BOM & Security Benchmarks Maintain a live inventory of models, scored across security, safety, and business alignment ⚡ MCP Server Discovery & Scanning Automatically identify and scan MCP servers for vulnerabilities ⚡ Compliance-Ready Reporting Generate reports aligned with OWASP LLM Top 10, EU AI Act, and more With AI Asset Management, enterprises finally gain the visibility and control to scale agentic AI with confidence. Learn more >>> https://lnkd.in/e8ucipph Kristian Kamber Ante Gojsalic Michael Sutton Chenxi Wang, Ph.D. David Endler Karol Lasota Stanislav Sirakov Petar Tsachev Lars Godejord Jure Mikuž Manoj Apte Joseph Thacker Sergej Epp Daniel Miessler 🛡️ John Stewart Ofer Ben-Noon Saša Zdjelar Julie Tsai

I’m proud to announce 𝗦𝗣𝗟𝗫 𝗔𝗜 𝗔𝘀𝘀𝗲𝘁𝘀! It all started a year ago with idea of 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗥𝗮𝗱𝗮𝗿, open-source tool created to improve precision of Agentic AI Red Teaming. Agentic Radar evolved into an enterprise-ready discovery module that reveals vulnerabilities within AI BOM and AI Workflows living in enterprise source-code infrastructure. Why this was crucial for SPLX platform? 🔎 We finally closed an AI-SPM loop that starts with 𝗱𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆 𝗼𝗳 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀, 𝗟𝗟𝗠 𝗺𝗼𝗱𝗲𝗹𝘀, 𝗠𝗖𝗣 𝗦𝗲𝗿𝘃𝗲𝗿 𝗮𝗻𝗱 𝗔𝗴𝗲𝗻𝘁 𝘁𝗼𝗼𝗹𝘀 in your environment. 🧪 After all repositories are scanned and potentially vulnerable AI components are identified, we provide one-click 𝘁𝗿𝗮𝗻𝘀𝗶𝘁𝗶𝗼𝗻 𝘁𝗼 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗲𝗱 𝗔𝗜 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗧𝗲𝘀𝘁𝗶𝗻𝗴. 🛡️ Last part is an 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗲𝗱 𝗿𝗶𝘀𝗸 𝗿𝗲𝗺𝗲𝗱𝗶𝗮𝘁𝗶𝗼𝗻 by hardening your business logic and dynamically generating input/output filtering rules for your guardrails. While having it in Alpha version, we already collected 20+ feature requests for AI Assets, so step-by-step it should get as rich as other parts of SPLX platform. Congrats to whole SPLX team and our design partners on this huge milestone! 👇 Link for more details is in the comment.

After a great experience at BSides Frankfurt, I'm happy to announce probably my last workshop for this year.. Join me at Security BSides Kraków to unpack the real-world risks behind today’s GenAI deployments. 𝗘𝘃𝗮𝗱𝗶𝗻𝗴 𝗚𝗲𝗻𝗔𝗜 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗗𝗲𝗳𝗲𝗻𝘀𝗲𝘀 🕤 Sept 27 | 09:45–10:30 📍 BSides Krakow In this session, we’ll break down: - The new attack surface GenAI creates - The most common security missteps teams are making today - What to prioritize in GenAI risk assessments, and why If you're red teaming, securing, or building with LLMs, this session is for you. Come challenge your assumptions and break things, safely. 🎟️ Link to tickets in the comments