AccuKnox Named Top AI Security Startup at Security BSides Bangalore 2025
'Innovator and Leader in AI Security'
AccuKnox was given the #1 AI Security Startup of 2025 Award by EmergeX: Unveiling Innovation at the influential BSides Bangalore Cybersecurity Conference in Bangalore. AccuKnox presented its Agentic AI runtime security solution. The competition judges were very accomplished CyberSecurity leaders and voted AccuKnox as the leading innovator. The unique and highly differentiated innovation in the area of AI Agentic Security positions AccuKnox ahead of established industry incumbents in addressing a vexing problem in AI Technology Adoption.
The recognition positions AccuKnox as the platform leader and innovator poised ahead of established incumbents who offer legacy/pathwork solutions to this complex problem.
AccuKnox presented clear user problems when it comes to security issues faced by users wanting to deploy Agentic AI solutions, such as:
Automated Red Teaming to understand the LLM guardrails posture
Providing visibility into AI pipelines and helping users to identify Shadow AI
Securing untrusted models at Runtime using AccuKnox's patented/developed Sandboxing engine
Handling PredML as well as GenAI as part of the same platform
Securing AI Infra, Apps, and AI Models/Datasets with full Enterprise integration options
The following is a summary of the AccuKnox Agent AI solution
Sandboxing Agentic AI deployments
Securing datasets leveraged by RAGs and model fine-tuning
Automated Red teaming of LLMs
AI Security Posture Management for managed and unmanaged deployment
A real-time demo was showcased that showed how Guardrails of the popular LLM (Anthropic Claude) can be bypassed by simple prompt engineering and how the AccuKnox solution could be used to prevent the attacks at the next level because of sandboxing.
AccuKnox AI Security solutions offer the following unique differentiators:
During the EmergeX demo, they showcased how easy the asset onboarding process is, which immediately provides a consolidated view to the users about their AI Security Posture.
The Agentic AI solution not only talks about Runtime Prompt Firewall but also about sandboxing the Agentic AI solution.
There was a question from one of the judges about the licensing model. AccuKnox's licensing model is based on the number of deployed models, which is easy to quantify.
AccuKnox presentation at BSides Bangalore can be viewed here
Supporting Quotes
" Agentic AI isn't just about automation—it's about intelligent delegation. In a world overwhelmed by complexity, it marks the rise of proactive digital partners that think, plan, and act alongside us", Golan Ben-Oni, CIO, IDT Telecom
Congratulations to the AccuKnox team for winning the EmergeX: Unveiling Innovation contest! Your demonstration of cutting-edge AI-driven security solutions showcases the transformative potential of agentic AI across the industry,' said Sujatha Yakasiri, Founder - Security BSides Bangalore and W3-CS (Worldwide Women in Cybersecurity)
' AccuKnox's win at BSides Bangalore is more than a trophy—it's a validation of years of deep research, customer obsession, and fearless innovation. As enterprise buyers shift from bulky, outdated tools to nimble, AI-enhanced platforms, AccuKnox stands at the forefront of this wave. We didn't build this for awards—we built it for the enterprise teams stuck with legacy tech that's failing them,' said Rahul Jadav, co-founder, CTO, AccuKnox. 'Winning EmergeX is incredible validation, but we derive immense satisfaction from the fact that clients can adopt AI-technology and deliver shareholder value and do it most safely and securely.'
AccuKnox is a next-generation CNAPP and Zero Trust security platform purpose-built for the cloud-native era. With its roots in open source and research from Stanford, AccuKnox delivers AI-enhanced detection, automated remediation, and seamless policy enforcement to secure enterprise workloads across Kubernetes, VMs, and multi-cloud environments.
Contact
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
4 hours ago
- Yahoo
Skyflow Unveils MCP Data Security for Enterprises and SaaS Companies
New MCP Data Security Platform Addresses Risks in Adoption of Agentic AI PALO ALTO, Calif., August 01, 2025--(BUSINESS WIRE)--Enterprise AI agents need access to customer data to deliver value. But each MCP connection adds a security risk. As the Model Context Protocol (MCP) becomes the backbone of agentic AI systems, Skyflow is launching its MCP Data Protection Layer—purpose-built for SaaS platforms and enterprises adopting MCP. MCP standard was introduced by Anthropic, and is now supported by OpenAI, AWS, and Google. It streamlines how AI agents connect securely to real-world tools like databases, SaaS tools and apps, etc. without writing custom code. But this introduces a new risk frontier: sensitive data like PII, PHI, and financial records flowing through the MCP servers without proper safeguards. Skyflow Addresses Sensitive Data Protection Challenges with MCP Use Unlike traditional DLP tools that simply block data, Skyflow takes a more intelligent approach. Its unique polymorphic data protection engine dynamically transforms sensitive information in real time by masking, tokenizing, or rehydrating fields based on policy and user permissions. This ensures security and compliance without breaking AI agent functionality. Skyflow MCP Data Protection is available in two deployment models: Skyflow MCP Gateway: A proxy layer, which can be integrated into existing proxy servers that sits between MCP servers or agents and backend data sources, enforcing field-level privacy policies without requiring application changes. Skyflow MCP Server SDK: An embeddable library that developers can use to build privacy controls directly into MCP server implementations and agentic apps. Both options include enterprise-grade privacy features: Use case-aware redaction and de-identification Entity-preserving transformations for agent reasoning Contextual rehydration for authorized users Secure memory handling to prevent PII retention Full audit trails for GDPR, HIPAA, and other regulatory compliance "As AI agents start connecting to more real-world data through MCP, companies need privacy infrastructure that can keep up," said Anshu Sharma, CEO of Skyflow. "Skyflow helps developers and SaaS platforms protect sensitive data without slowing down AI workflows—making secure, compliant AI deployment possible at scale." Enterprises and SaaS companies across retail, financial services, healthcare, travel and hospitality can deploy Skyflow's MCP protection to enable AI agents with sensitive data access while maintaining regulatory compliance. Today's MCP Data Protection offering announcement extends Skyflow's AI security roadmap. It builds on the earlier launch of Agentic AI Security and Privacy Layer launched last year, and the GPT Privacy Vault introduced in 2023. To understand the specific privacy risks MCP servers introduce, read our detailed blog post: Building Secure AI Agent Architecture with Model Context Protocol About Skyflow Skyflow is the security and privacy platform for the modern AI data stack built to radically simplify how companies isolate, protect, and govern their customers' most sensitive data. With its Data Privacy Vault, Skyflow enables businesses to store, process, and share sensitive data securely. Leading investors back Skyflow, and the company is trusted by Fortune 500 and growth companies across financial services, healthcare, travel & hospitality, and retail. View source version on Contacts Media Contact: Rajsi Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Forbes
6 hours ago
- Forbes
Agentic AI: How Organizations Can Prepare Now For Tomorrow's Unknowns
Rob Mason is the CTO at Applause, helping leading brands quickly release high-quality digital experiences. In an earlier article on functional testing, I outlined what I saw (at the time) as the major challenges facing organizations looking to push out high-quality digital products while keeping their costs down. As I noted then, 'The common thread here is that the technology and the market are always evolving. Products are getting more complicated with so many different dynamic elements.' And I concluded, '[A]That was May 2024. In that month alone, OpenAI released GPT-4o, Microsoft introduced Copilot + PCs and the U.S. Senate rolled out its new AI policy roadmap. I was right about the evolution, but I continue to be astounded by the pace. If generative AI was the buzzword of the early 2020s, then agentic AI is certainly on track to take the next half of this decade by storm. The agentic AI market is expected to reach $126.89 billion by 2029. Gartner predicts that 33% of enterprise software applications will incorporate agentic AI by 2028, and that 15% of day-to-day work decisions will be made by this technology by the same time. But how do we define agentic AI? And how can organizations prepare for it? What is agentic AI? While there are several articles on this subject, here is a straightforward summary. Traditional AI is rules-based and uses static and structured data to execute predictable tasks. Generative AI creates outputs based on input or training data you feed it—think ChatGPT creating summaries or DALL-E creating images. Agentic AI is different because it is designed to make decisions. It works with both structured and unstructured data and adapts, taking action based on context, feedback and/or a change in objective. Benchmarks show that agentic AI systems are able to complete up to 12 times more multi-step tasks than standard large language models. Its capabilities are astounding, as are its challenges. Because agentic AI is dynamic, transparency and trust are key to understanding why the AI makes certain decisions. Personalization also means there's a chance for it to become unpredictable or inconsistent, as well as unsympathetic and just generally unhelpful. The risks are also higher because we have AI touching all aspects of business—every piece of infrastructure, every department, every audience, including customers, employees, APIs and beyond. Essentially, the astronomical number of dimensions that need to be tested (a phrase I used way back in 2024) is growing exponentially. Tomorrow's dimensions are unknown. Traditional testing approaches will not work for agentic AI or whatever new technology comes next. The dimensions are too big and complex. Evolution is constant, so testing needs to measure not just output, but also how the AI interacts with people's (i.e., real humans') emotions and intentions. Establishing trust from the beginning by prioritizing transparency, user feedback and responsible design means that AI can act as a partner, not just a tool that solves today's problems. To get started, organizations need to ensure that they can scale scenario-based agentic AI testing to evaluate all of the environments where these agents might operate. They should look at real-world workflows and see if the AI can handle multi-step tasks or a change in plans. For example, test how an agent handles booking a flight while adjusting smart home devices in response to last-minute travel changes. Or, see how it does with shopping and procurement—ask the agent to find and purchase an item, charge the correct payment method and successfully complete the task. Organizations also need to incorporate governance and clear escalation paths from the start to help ensure that agentic AI is responsible AI, ready to adapt if things get too risky. Transparency, in particular, is key to this next phase of AI. Users need to understand why your AI is taking action, so make sure that they are able to review, reverse or change their minds if needed. On the back end, this means comprehensive logging to help make every decision auditable—a change for QA teams who were previously able to do their jobs without seeing detailed debug logs. And, for all the hype about AI taking away jobs, collaboration across multiple job functions is going to be required to make sure that AI is actually useful. AI is going to work across product, IT, compliance, customer experience, HR and marketing. Feedback from all of those stakeholders will be key to its success. Finally, organizations can establish clear success metrics now while still exploring agentic AI use cases. AI hype is real, and sometimes taking a step back to just ask, 'What is the business problem we're trying to solve?' can help level set. If that's not clear, it's time to take a breath. Once that's established, you can decide which metrics—accuracy, response time, customer experience scores, etc.—make the most sense for your business. When in doubt, think about the humans. Agentic AI aggravates so many of the existing complexities of modern testing that it's hard to know where to start. Crowdtesting, training with diverse datasets and red teaming are all best practices that I've highlighted in the past. Over the next few articles, as agentic AI technology continues to evolve and become more pervasive, I plan to dive into specific ways organizations can think about testing and measurement when dimensions are so fluid. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Entrepreneur
12 hours ago
- Entrepreneur
Cybersecurity Leader SAFE Raises USD 70 Mn in Series C Funding
The company also introduced a major upgrade to its Cyber Risk Singularity platform — the world's first fully autonomous Continuous Threat Exposure Management (CTEM) solution powered by Agentic AI. You're reading Entrepreneur India, an international franchise of Entrepreneur Media. SAFE, a leader in Cyber Risk Quantification (CRQ) and autonomous Third-Party Risk Management (TPRM), has raised USD 70 million in a Series C funding round led by Avataar Ventures. Other participants include Susquehanna Asia Venture Capital, NextEquity Partners, Prosperity7 Ventures, and existing investors such as Eight Roads, John Chambers, and Sorenson Capital. The funds will be deployed to strengthen SAFE's position in the cyber risk management market and to advance the development of Agentic AI-based reasoning models aimed at achieving CyberAGI. The company also introduced a major upgrade to its Cyber Risk Singularity platform — the world's first fully autonomous Continuous Threat Exposure Management (CTEM) solution powered by Agentic AI. "This is a defining moment in our pursuit of CyberAGI," said Saket Modi, Co-founder and Chief Executive Officer of SAFE. "When we launched our platform in 2020, we focused on Cyber Risk Quantification as the foundation of cyber risk management. Since then, we have expanded into autonomous Third-Party Risk Management and now into Continuous Threat Exposure Management, each critical to achieving our mission." SAFE's CTEM solution is designed to address limitations in traditional tools, which often rely on black-box scoring models and static data aggregation. By leveraging multiple autonomous AI agents, the platform aims to provide transparent and actionable risk insights. Founded in 2012 by Saket Modi, Viditkumar Baxi, and Rahul Tyagi, SAFE originated from an incubation program at the Indian Institute of Technology Bombay. Headquartered in Palo Alto, California, the company offers cybersecurity risk quantification and management services, enabling organisations to continuously measure, prioritise, and mitigate cyber risks across their digital infrastructure. SAFE's platform covers CRQ, CTEM, and TPRM, making it one of the few companies to integrate these capabilities into a single system. The firm counts major corporations including Google, Fidelity, T-Mobile, Chevron, and IHG among its clients. Nishant Rao, Founding Partner at Avataar Ventures, noted, "Cybersecurity is now a boardroom and CEO-level priority. What differentiates SAFE is its role as a strategic intelligence layer across the cybersecurity stack, supported by a strong execution track record and consistent triple-digit growth since 2020." With this funding, SAFE's total capital raised surpasses USD 170 million. The company intends to continue expanding its capabilities, building AI-native models, and reinforcing its leadership in autonomous cyber risk management solutions.
