Ambuj Kumar, CEO of Simbian.AI, shares how AI agents are transforming cybersecurity with precision, efficiency, and innovation.
Hello Ambuj, it’s a wonderful opportunity to have you with us. Could you please tell us about your professional journey from co-founding Fortanix to co-founding Simbian.AI?
Thank you for having me. My journey from Fortanix to Simbian.AI has been exciting and rewarding. At Fortanix, we focused on confidential computing and security, pushing the boundaries of data protection. With Simbian.AI, we are now taking that vision a step further by using AI to automate security tasks. This means that organizations can now protect themselves using a suite of AI agents trained specifically for cybersecurity tasks, which perform autonomously. We’ve raised about $10 million in funding, and it’s been an incredible journey so far.
Every vendor in the market offers generative AI solutions, but there are nuances when it comes to practical application. How should organizations choose a vendor for these solutions?
That’s a great question. When choosing a vendor, the most important thing to ask is whether the solution is just a cool technology or if it will genuinely impact your top or bottom line. At Simbian, we focus on solutions that actually make our customers’ lives easier and their organizations more secure. For example, if you run a Security Operations Center (SOC), our AI can help you cover more security events with fewer resources, making your life easier and keeping your company safe. Another key factor is ensuring that the AI has guardrails in place to avoid hallucinations or mistakes—before deploying any AI into production.
What strategies have you employed at Simbian AI to provide effective solutions for the complex issues of cybersecurity?
Cybersecurity is a complex and evolving field, and Simbian prides itself on creating innovative, first-of-its-kind solutions. For instance, we built the industry’s first trusted LLM, a large language model specifically trained for security, which doesn’t hallucinate and remains under control at all times. We also pioneered AI agents for cybersecurity, which can take very complex actions autonomously. Our AI agents help SOC teams by eliminating more than 90% of their routine tasks, allowing them to focus on more critical issues.
Considering the rapid growth of AI in cybersecurity, what are some ways to optimize its operations?
AI can be used in a variety of cybersecurity applications—from triaging alerts and threat hunting to incident response and GRC tasks. One important aspect to keep in mind is how different AI applications work together. You don’t want to end up with a set of disparate AI tools that don’t communicate effectively. Instead, I recommend a platform-based approach, like we offer at Simbian, where all applications work cohesively. This helps in streamlining operations and prevents the common complaint from security leaders about managing too many different tools.
How do you define the roles of co-pilots and agents within the context of generative AI?
In the context of generative AI, co-pilots are tools that assist the user in decision-making by offering guidance or suggestions. They work alongside you to complete tasks. Agents, on the other hand, are more autonomous. They don’t just assist—they act independently and take complex actions based on the data they process. Simbian’s AI agents are an example of this autonomy, making cybersecurity operations faster and more efficient.
What are the key differences between a generative AI co-pilot and an AI agent?
The key difference lies in their level of autonomy. Co-pilots provide suggestions and help users complete tasks in real-time. AI agents, however, are autonomous—they don’t need continuous user input. They can execute tasks on their own, which is particularly useful in environments like cybersecurity where fast, independent decisions are required. This is exactly what we offer with Simbian’s GRC AI Agent and SOC AI Agent.
What do you think might be the potential risks with the autonomous operations of AI agents?
One of the biggest risks with autonomous AI agents is the potential for hallucinations—when the AI generates incorrect or fabricated information. This is why it’s critical to have strong guardrails in place to prevent errors from leaving the AI system. In cybersecurity, precision is key, and we can’t afford for AI to make mistakes. At Simbian, we’ve trained our models to be secure and reliable, ensuring that mistakes are either not made or caught before they become an issue.
How would you define ‘hallucinations’ when it comes to generative AI in cybersecurity?
Hallucinations in generative AI occur when the system creates new information that wasn’t in its training data. For instance, if an AI is trained on certain sentences and it generates a new sentence that wasn’t part of the original data, that’s a hallucination. While this creative ability is useful in some applications, it’s problematic when dealing with fact-based questions. In cybersecurity, hallucinations can lead to misinformation, which is unacceptable. So, it’s crucial that AI models like ours at Simbian are trained to avoid this issue.
Could you share your insights on the future of AI in cybersecurity and provide some advice for organizations looking to grow as trusted LLM providers?
The future of AI in cybersecurity is very promising. AI adoption is accelerating due to increasing risks and the pressure to manage more complex tools. My advice for trusted LLM providers is to focus on building a platform that allows security teams to do more with fewer resources. At Simbian, we offer that kind of integrated, resource-efficient platform for SOC and GRC teams.
Finally, could you share your favorite quote and its relevance to professional life?
One of my favorite quotes is from Yeats: “In dreams begin responsibility.” It reminds me that while we’re free to dream and aspire, we also need to take on the responsibilities that come with those dreams. Achieving great things requires hard work and commitment. At Simbian, we apply this mindset by not just dreaming of revolutionizing cybersecurity with AI, but also taking on the responsibility to deliver highly reliable, autonomous AI agents for next-gen security. It’s a vision that drives our team to keep pushing the boundaries of what’s possible in cybersecurity.
Ambuj Kumar
CEO and Co-Founder of Simbian
Ambuj is co-founder and CEO of Simbian, which is using AI to automate security. Previously, he was co-founder and CEO of Fortanix where he created a new security category, Confidential Computing. He has raised $150m+ across the two companies and is a tech visionary, with more than 100 patents to his name. Ambuj started his career as lead engineer at NVIDIA for 8 years where he built the first few generations of GPUs, powering today’s AI revolution. Ambuj has been recognized with dozens of awards for his contribution to technology and business. He holds an M.S. degree from Stanford and a B.Tech from IIT Kanpur, India (Gold Medalist).
