Saint John-based TrojAI has announced it has partnered with New York cybersecurity firm Wiz as part of the American company’s release of its new AI Application Protection Platform, or AI-APP.

Founded by James Stewart and Stephen Goddard in 2019, TrojAI sells cybersecurity software to guard artificial intelligence systems against having malicious code injected via training data. The company is now helmed by CEO Lee Weiner, who was appointed in 2024.

Wiz announced the new product last week, saying it’s aimed at helping organizations identify and manage risks in artificial intelligence systems. TrojAI’s role in the new product is to focus on testing AI systems for vulnerabilities.

“TrojAI and Wiz integrate to bring AI red teaming findings from TrojAI into Wiz, revealing vulnerabilities like jailbreaks, prompt injection exploits, and data leakage from your AI Endpoints,” said Wiz on its website. “Wiz enriches these findings with cloud context to surface attack paths to your AI application, helping teams prioritize exploitable risks.”

The companies say AI-related risks are often spread across multiple layers, including models, software tools, data, and underlying infrastructure. This makes it difficult for security teams to detect and prioritize threats using traditional methods.

TrojAI, which raised $7.8 million in venture capital in 2024, specializes in “red teaming,” a process where systems are deliberately tested to uncover weaknesses. In this case, it identifies issues such as prompt injection (manipulating AI inputs to produce unintended outputs), system “jailbreaks,” and potential data leaks.

Wiz’s platform then adds context from cloud environments and live system operations to these findings. This combined view is designed to help organisations understand how individual vulnerabilities could be exploited in real-world scenarios, rather than treating them as isolated risks.

The companies say this approach allows security teams to focus on the most significant threats – those that are both technically feasible and likely to be targeted – rather than theoretical concerns.

The announcement reflects growing industry attention on securing AI systems as their use expands across business operations. By combining vulnerability testing with broader system analysis, Wiz and TrojAI aim to provide organizations with clearer guidance on how to protect AI applications in practice.