At People.ai, we are committed to pushing the boundaries of AI revenue platform modernization with built-in generative AI capabilities. As enterprises increasingly rely on AI to optimize sales data ingestion, filtering, matching, and enrichment within their CRMs, security remains a top priority. That’s why we partnered with Google and Mandiant, leaders in penetration testing and cybersecurity.
This collaboration ensures that our generative AI technology is not only cutting-edge but also built on a foundation of robust security measures to protect our customers.
Strengthening Generative AI Security with Google | Mandiant’s Expertise
Our partnership with Google and Mandiant was focused on conducting a comprehensive security assessment of our Generative AI models and Conversation Intelligence agent. The primary goal was to identify and mitigate security risks across various attack vectors, including Cross-Site Scripting (XSS), HTML, SQL, server-side injections, and access control vulnerabilities.
A Security-Focused Approach
Mandiant’s assessment covered multiple critical security areas, ensuring a well-rounded evaluation:
- Discovery – Analyzed People.ai’s platform functionality, design principles, and potential attack surfaces.
- Configuration Management – Evaluated our web, application, and database services to ensure up-to-date security patches and configurations.
- Authentication – Verified user identity validation before granting access to restricted functionalities and data.
- Authorization – Ensured proper access controls, restricting users to only their authorized data and functionalities.
- Data Validation & Protection – Assessed our LLMs' ability to handle user inputs securely and verified encryption protocols for data storage and transmission.
Key Findings and Outcomes
Through this rigorous assessment, we:
- Simulated various attack scenarios, including email ingestion mismatches and injection attacks (XSS, SQL, etc.).
- Reviewed all running services, checking for outdated versions, libraries, and frameworks.
- Evaluated authentication mechanisms to ensure resilience against tampering.
- Investigated potential unauthorized data access across customer datasets.
Building Trust Through Security
The completion of this extensive security evaluation reaffirmed our commitment to data security and trust. James Wedewer, Director of Information Security at People.ai, emphasized, “Meeting security fundamentals in generative AI systems is crucial to maintaining the confidentiality and integrity of our customer data. Our partnership with Google and Mandiant allowed us to achieve this through rigorous testing and validation.”
As we continue to enhance our AI-driven data platform, our mission remains clear: to provide enterprises with secure and trustworthy generative AI solutions. Partnering with industry leaders like Google and Mandiant ensures that our customers can confidently leverage AI, knowing that Trust and Safety are at the core of our innovation.