TLDR: Lanai Inc. has launched a new edge-based AI Observability Agent designed to detect and mitigate ‘shadow AI’ usage within organizations. This innovative platform runs detection models directly on corporate devices, offering real-time visibility into employee generative AI interactions across all applications without sending sensitive data outside the company’s network. The move aims to address the growing risks of unauthorized AI tool usage and potential data leaks, a problem that a LayerX Ltd. study indicates makes 89% of enterprise AI usage invisible to IT and security teams.
San Francisco, CA – September 10, 2025 – AI monitoring startup Lanai Inc. today announced the release of its groundbreaking edge-based AI Observability Agent, a solution poised to revolutionize AI governance by tackling the pervasive issue of ‘shadow AI’ within enterprises. The platform distinguishes itself by deploying lightweight AI detection models directly onto corporate devices, circumventing the need to route sensitive conversations through centralized cloud infrastructure. This approach provides unparalleled, prompt-level visibility into employee interactions with generative AI across any application, whether embedded, native, or newly released, all while ensuring data remains within the organization’s network.
According to Lanai, ‘shadow AI’ has become a critical challenge for businesses. A recent study by LayerX Ltd., cited by the company, reveals that a staggering 89% of all enterprise AI usage remains completely invisible to IT and security teams. This invisibility stems from various employee behaviors, including the use of personal ChatGPT accounts for work-related tasks, software engineers leveraging unauthorized coding tools like Codeium and Cursor for faster development, and employees inadvertently feeding sensitive information into AI features embedded within popular SaaS platforms such as Salesforce Einstein, Microsoft Copilot, Adobe Firefly, Slack AI, HubSpot AI, Notion AI, and Figma AI.
Lexi Reese, co-founder and Chief Executive of Lanai, emphasized the severity of the problem: “The shadow AI problem is exploding as employees are using personal accounts, unapproved coding agents and embedded AI features in everyday SaaS tools, and this all sits outside IT’s visibility. Traditional tools might catch someone visiting ChatGPT.com, but they have no idea if that employee just shared company trade secrets or a client’s sensitive revenue information.” Reese further explained that Lanai’s agent embeds these observability models directly onto each employee’s device, bringing this hidden activity to light and establishing guardrails to prevent sensitive information from being shared with third-party AI, thereby ensuring safety without compromising speed or innovation.
Steve Herrod, Lanai co-founder and former VMware CTO, likened the shift to a fundamental change in monitoring philosophy. “We’re essentially moving AI observability from the network to the edge,” Herrod stated. “It’s like the shift from monitoring server rooms to having telemetry inside every virtual machine. Traditional approaches see network traffic or ping static lists that do not update dynamically; we see the actual prompt interactions and where real risks and value live.”
Also Read:
- SPLX Unveils AI Asset Management Solution for Comprehensive Enterprise AI Stack Security
- OneTrust Unveils AI-Ready Governance Innovations Amid Surging AI Risk Management Demands
Deployment of Lanai’s AI Observability Agent is remarkably swift, taking less than 24 hours via standard Mobile Device Management (MDM) systems, with no infrastructure changes required. The platform offers dynamic detection across any application, eliminating the need for static lists, and performs real-time prompt and response analysis to prevent sensitive data or workflow insights from being leaked. This on-device processing ensures complete data sovereignty. Early deployments have already yielded significant results; in one instance, an information security team confident in their controls discovered 27 unauthorized AI tools in use within just four days of Lanai’s deployment. Other examples include a healthcare CISO uncovering 31 AI tools versus 5 approved, flagging sensitive patient data, and cutting $480K in licenses while boosting productivity by 40%. A FinTech CIO/CISO reported exposing deal terms in personal ChatGPT accounts and confidential data in unsanctioned SaaS AI, leading to $320K in savings. By providing intelligence on AI tool usage, Lanai empowers companies to approve beneficial services and shut down risky ones, transforming AI governance from a hindrance into an accelerator for innovation and productivity.
