TrojAI Detect Advances AI Red Teaming with Agentic and Multi-Turn Attacks

Enabling Enterprises to Automate Real-World Risk Assessments at Scale

SAINT JOHN, NB and BOSTON, July 29, 2025 /PRNewswire/ -- TrojAI, the enterprise security platform for artificial intelligence (AI), today announced significant innovations to its AI red teaming solution, TrojAI Detect, introducing support for agentic and multi-turn attacks. This makes TrojAI Detect the most advanced AI red teaming solution on the market, capable of simulating sophisticated adversarial attacks to uncover risks in AI models, applications and agents. These innovations will be on display next week at booth #5916 at the Black Hat USA 2025 conference.

As enterprises move from experimenting with AI to deploying applications and agents in production, the focus is shifting to assessing risk and protecting AI systems from threats. With growing concerns around prompt injection, data leakage and jailbreaking, organizations are demanding deeper visibility into AI model behavior to manage real-world risks at scale.

The latest release of TrojAI Detect enables security teams to simulate complex adversarial attacks, automating multi-turn and agentic red teaming techniques. This expanded coverage marks a leap forward in red teaming sophistication, allowing enterprises to test their AI with advanced, automated and dynamic workflows that mimic the way real-world adversaries operate.

"These new capabilities reflect an important step forward in how we assess and understand the behavior of AI systems," said Lee Weiner, CEO of TrojAI. "With agentic and multi-turn attack types, we're moving from single-shot probes to persistent, context-aware adversarial agents. It's the most advanced form of behavioral testing available, and it brings our customers closer to continuous, autonomous AI assurance."

TrojAI Detect leverages new agentic and multi-turn techniques to enable enterprises to automate real-world attacks for deeper understanding of agent and model behavior that includes state and history. These automated attacks include both dynamically and computationally generated prompts designed to uncover behavioral vulnerabilities across diverse AI architectures. New attack types include the following:

• Agentic Attacker: Finds jailbreaks using a coordinated multi-agent approach
• Conversation Obfuscation: Hides malicious intent across multiple prompts
• Undesirable Content: Uses LLMs to elicit toxic or unwanted content

To learn more about the support for agentic and multi-turn attacks, read the full blog.

About TrojAI
TrojAI's mission is to enable the secure rollout of AI in the enterprise. TrojAI delivers a comprehensive security platform for AI. The best-in-class platform empowers enterprises to safeguard AI models, applications and agents both at build time and run time. TrojAI Detect automatically red teams AI models, safeguarding model behavior and delivering remediation guidance at build time. TrojAI Defend is an AI application and agent firewall that protects enterprises from real-time threats at run time. By assessing the risk of AI model behavior during the model development lifecycle and protecting it at run time, TrojAI delivers comprehensive security for AI models, applications and agents.

Media Contact
[email protected]

View original content to download multimedia:https://www.prnewswire.com/news-releases/trojai-detect-advances-ai-red-teaming-with-agentic-and-multi-turn-attacks-302515382.html

SOURCE TrojAI