ASAPP Launches Continuous Red Teaming for Enterprise AI Security
ASAPP has introduced Continuous Red Teaming, an adversarial AI testing capability built on Promptfoo that automatically screens for over 50 vulnerability types during development.
ASAPP has launched Continuous Red Teaming, a new capability that integrates adversarial AI testing directly into its model evaluation framework. The feature is built on Promptfoo, an AI security platform that helps enterprises detect and address vulnerabilities in AI systems during development. Promptfoo continuously runs automated tests across ASAPP's AI systems, screening for more than 50 vulnerability types to give enterprise customers the real-time data they need to trust their AI in production.
The new capability addresses a growing concern in enterprise AI adoption: the risk of deploying models with undetected security flaws. Traditional red teaming often occurs as a one-time exercise before deployment, leaving gaps as models are updated or exposed to new data. ASAPP's approach shifts to continuous testing, embedding security checks throughout the development lifecycle rather than treating them as a final gate.
Continuous Red Teaming covers a broad range of vulnerability categories, including prompt injection, jailbreaking, data leakage, and model inversion attacks. By automating these tests, ASAPP aims to reduce the manual effort required for security teams to keep pace with rapid AI development cycles. The integration with Promptfoo provides a standardized testing framework that can be applied consistently across different AI models and use cases.
For enterprise customers, the benefit is real-time visibility into AI security posture. The system generates alerts and detailed reports when vulnerabilities are detected, allowing developers to remediate issues before models reach production. This aligns with industry best practices for secure AI development, such as those outlined in the NIST AI Risk Management Framework and OWASP's Top 10 for LLM Applications.
The launch comes amid a broader industry push to secure AI systems. Recent incidents, including data exfiltration via prompt injection and model poisoning attacks, have highlighted the need for proactive security measures. ASAPP's offering positions it alongside other vendors like Microsoft, which recently open-sourced its RAMPART and Clarity tools for AI agent security, and NetSPI, which launched an AI-powered continuous pentesting platform.
ASAPP's focus on enterprise AI systems is notable given the company's specialization in customer experience and contact center AI. Its models handle sensitive customer data, making security a critical requirement for adoption. By embedding red teaming into its development pipeline, ASAPP aims to build trust with customers who may be hesitant to deploy AI in production due to security concerns.
The Promptfoo platform, which powers the new capability, is an open-source tool that has gained traction in the AI security community. It supports testing for a wide range of attack vectors and integrates with popular AI frameworks. ASAPP's adoption of Promptfoo underscores the growing trend of leveraging open-source security tools to build enterprise-grade defenses.
As AI adoption accelerates across industries, the demand for automated security testing solutions is expected to grow. ASAPP's Continuous Red Teaming represents a step toward making adversarial testing a standard part of the AI development lifecycle, rather than an afterthought. The capability is available now to ASAPP's enterprise customers.