Noma provides security from business AI data to deployment

As 2024 draws to a close, the state of business technology is that companies of all sizes are eager to leverage their data into generative AI applications that improve internal (employee-centric) or external (customer/partner-centric) processes.

Ensuring they do this securely, however, is another challenge, especially for companies that don’t specialize in security. For many enterprises, their existing security solutions may also be inadequate or unprepared for the AI ​​era and the many possibilities they want to unleash with their AI-channeled data.

Enter Nomaan Israeli startup specialized in AI enterprise security, which today leaves stealth mode with a Series A round led by $32 million Ballistic Ventures and backed by Glilot Capital Partners and Cyber ​​Club London, as well as angels including the Chief Information Security Officers (CISOs) from companies such as McDonald’s, Google DeepMind, Twitter, Atlassian, BNP Paribas, T-Mobile and Nielsen.

Noma provides a comprehensive security platform that ensures the integrity of enterprise customers’ data from the very beginning, before they do anything with it, through to using it to train and/or deploy AI models and custom applications.

The platform is already used by several Fortune 500 companies.

Addressing security challenges in the data and AI landscape

Niv Braun, co-founder and CEO of Noma, told VentureBeat in an interview about the urgent need for targeted security in AI workflows. “

“Today’s AI and data science models face unique security risks, such as rapid injection and data breaches, that are simply not covered by standard security tools,” he says.

These issues are becoming increasingly common as organizations experience security incidents due to misconfigured MLOps tools and unverified open source models.

This gap inspired Braun and his co-founder, Alon Tron, to create Noma.

“My co-founder Alon and I served in the military together, and we both saw firsthand the gap in security tools for data science and AI workflows,” said Braun. “In application security, we had tools that helped software engineers work securely, but for data teams – data scientists, engineers and analysts – there was nothing comparable. They were left unprotected.”

Both co-founders served in Israel’s elite intelligence unit 8200. Combining expertise from their backgrounds in security and data science, they quickly formed a team versed in AI and application security.

What Noma’s three-tiered platform offers

Noma’s platform is designed to protect every stage of AI model development and operation, and includes security tools covering:

1. Data and AI supply chain security: Ensures secure environments, pipelines, and development tools, mitigating the risk of compromised data and AI supply chains.
2. AI Security Posture Management (AI-SPM): Provides a comprehensive inventory and security management solution for both first- and third-party AI models, with the goal of protecting assets before they go into production.
3. AI threat detection and response: Actively monitors AI applications to detect hostile attacks in real-time and enforce security protocols at runtime.

Braun highlighted the consolidation that Noma’s platform offers customers. “Our platform includes three products: data and AI supply chain security, AI security posture management and AI runtime defense.”

But for those who wish: each of the three domains can be applied ad hoc, à la carte.

“A major strength of our platform is that it consolidates everything into one solution,” Braun explains. “Although customers can choose just one component, most prefer the comprehensive approach.”

Braun clarified that Noma offers the choice between a comprehensive enterprise license and a modular, product-based option, both based on an annual Software-as-a-Service (SaaS) subscription. He said 95% of customers have chosen the integrated, all-in-one approach so far.

Braun’s comments demonstrate that the enterprise license is positioned as the most cost-effective, flexible choice for customers seeking comprehensive, organization-wide access to Noma’s solutions.

Maximum flexibility and ease of use

Noma’s platform is compatible with a variety of environments, supports cloud-based, SaaS or self-hosted configurations and can be installed in minutes with no code changes required.

“Integration is easy,” says Braun. “All customers have to do is connect our platform via API, after which we automatically map and scan everything in their environment.”

This frictionless setup means data science teams can implement security controls without disrupting their workflows, a feature that Noma highlights as essential for rapid, AI-powered developments.

Kobi Samboursky, Founder and Managing Partner at Glilot Capital Partners, praised the value of Noma’s unified approach in a press release: “AppSec has evolved over decades to include fragmented tools for static and dynamic analysis, open source, supply chain and runtime. Security teams have come to realize that they need consolidated solutions. Noma is uniquely positioned to tackle this problem from the ground up, consolidating multiple use cases into a single platform.”

Furthermore, Noma can be adopted by people without extensive training in security or data infrastructure.

“We work with data and AI teams as well as security teams, and our platform does not require deep expertise in either area,” he said. “Even in cases where security teams ran POCs (proof of concepts) without data science teams involved, they found it easy to integrate and use.”

At the same time, the platform turns these topics into digestible, easy-to-understand insights for employees working across all departments.

“The platform itself is very self-learning,” Braun noted. “It explains the basics of security in a way that application security teams are familiar with, but with a new ‘data and AI’ layer.”

Meeting the wishes and needs of the industry

As security and compliance become increasingly important in AI adoption, Noma aims to facilitate collaboration between data science and security teams.

“Our mission is to bridge the gap between data science and security teams, making it easy for both to work together to secure AI workflows,” said Braun.

Noma’s approach aims to improve transparency and simplify security processes.

“We make security simple for both teams, providing clear, understandable risk information and steps for remediation,” he added. “It’s all about reducing friction and improving collaboration.”

Jake Seid, co-founder and General Partner at Ballistic Ventures, emphasized the importance of safety from the start in a statement in a press release.

“As security and compliance become increasingly important for organizations adopting AI, embedding security from the start ensures innovation can flourish without compromise,” said Seid. “Noma’s approach gives AppSec teams complete visibility and confidence, while empowering data science teams to act quickly and drive business value.”

Noma’s ambitions are to lead the emerging field

Noma’s entry into the market marks an important step in securing AI-driven business operations at scale.

With the increasing use of AI in critical applications, the potential for security vulnerabilities in AI workflows becomes more acute.

Noma’s platform provides much-needed security, allowing businesses to leverage the potential of AI without compromising on security.

In addition, Noma actively contributes to AI security standards and has participated in the development of US government guidelines, such as NIST SP 800-218A, through its involvement in the OWASP AI Exchange.

With $32 million in new funding and early traction among high-profile customers, Noma aims to become a leader in the emerging field of data and AI lifecycle security.