Citadel AI Launches 'Lens Governance' to Unify AI Pre-deployment Review and Inventory Management
Citadel AI has begun offering a private beta for 'Lens Governance,' an AI governance tool that integrates pre-deployment review and software asset inventory management for generative AI. The tool facilitates collaboration between business departments (first line) and risk management departments (second line) by automating the review process. Through an inventory function that visualizes company-wide AI usage and risk distribution, it supports enterprises in safely and rapidly deploying and operating trustworthy AI.
📋 Article Processing Timeline
- 📰 Published: May 18, 2026 at 16:00
- 🔍 Collected: May 18, 2026 at 07:31
- 🤖 AI Analyzed: May 20, 2026 at 10:31 (50h 59m after Collected)
Citadel AI Inc. (CEO: Hironori Kobayashi) has announced the early-access launch of a private beta for 'Lens Governance,' a solution designed to unify the pre-deployment review of generative AI and the management of software asset inventories.
For businesses to fully and safely leverage generative AI, it is crucial for multiple departments—including application users, developers, and review boards—to collaborate. This requires establishing a rapid and appropriate risk review process, as well as building an inventory system to centrally manage risk assessments for all AI systems scattered throughout the organization.
'Lens Governance' is an AI governance tool that integrates these pre-deployment review procedures and inventory management, providing visibility into company-wide AI utilization and risk distribution. By consolidating complex management tasks into an intuitive and simple UI, it helps organizations overcome internal silos to promote the adoption and continuous operation of 'Trustworthy AI.'
## AI Governance as a Management Imperative
Generative AI technology is advancing so rapidly that not using it is no longer an option in business. Simultaneously, corporate legal and ethical responsibilities, which are directly linked to security and reputational risks, have reached unprecedented levels. How a company establishes enterprise-wide AI governance to actively utilize generative AI has become a critical management issue that can affect corporate value.
Evaluating generative AI is technically different and more challenging than for conventional systems. Establishing a cross-departmental review workflow and preparing test datasets requires enormous effort and time. As numerous AI applications and agent systems are introduced within companies, managing them all manually and maintaining audit trails is extremely difficult.
## Automating Governance Processes Based on Internal Roles
'Lens Governance' provides role-specific interfaces for application requesters (first line of defense, i.e., business units) and evaluators (second line, i.e., risk management and review departments). By controlling screen transitions according to each role's procedures, it enables a proper and swift AI governance review and approval process through collaboration.
Evaluators select appropriate datasets from the built-in 'Lens Safety Dataset' or the client's own datasets based on the application's risk level. The requester simply uploads the application's output based on this dataset, and the quality validation process proceeds automatically. The final approval is made by the evaluator, creating a 'Human in the Loop' configuration that balances automation with human oversight. The tool will also support verification based on legal systems and guidelines, such as Japan's AI Operator Guidelines.
Furthermore, the platform offers an inventory function (an 'AI asset ledger') to centrally manage and visualize which AI is running where and with what risk assessment. This feature also allows for scheduling periodic software asset audits and continuous quality assessments.
## Key Features of Lens Governance
- Automated approval processes and role-based screen transitions for evaluators (reviewers) and requesters (applicants).
- Inventory management to visualize AI utilization and risk distribution via a dashboard for management and evaluators.
- Scheduling for recurring and continuous validation, such as annual software asset audits.
- A built-in 'Lens Safety Dataset' covering 10 categories including safety, fairness, intellectual property, personal data leakage, and transparency for rapid and appropriate evaluation.
- Progressive support for datasets based on guidelines like the AI Operator Guidelines.
- Capability for custom evaluations based on client-specific metrics and use cases.
## About Citadel AI Inc.
Citadel AI is a global startup originating from Japan, focused on achieving the societal implementation of 'Trustworthy AI.' An international team, including engineers from Google's U.S. headquarters, leads development. The company provides proprietary AI governance tools that support the entire lifecycle of generative AI, from risk assessment at the implementation stage to guardrails during operation.
For businesses to fully and safely leverage generative AI, it is crucial for multiple departments—including application users, developers, and review boards—to collaborate. This requires establishing a rapid and appropriate risk review process, as well as building an inventory system to centrally manage risk assessments for all AI systems scattered throughout the organization.
'Lens Governance' is an AI governance tool that integrates these pre-deployment review procedures and inventory management, providing visibility into company-wide AI utilization and risk distribution. By consolidating complex management tasks into an intuitive and simple UI, it helps organizations overcome internal silos to promote the adoption and continuous operation of 'Trustworthy AI.'
## AI Governance as a Management Imperative
Generative AI technology is advancing so rapidly that not using it is no longer an option in business. Simultaneously, corporate legal and ethical responsibilities, which are directly linked to security and reputational risks, have reached unprecedented levels. How a company establishes enterprise-wide AI governance to actively utilize generative AI has become a critical management issue that can affect corporate value.
Evaluating generative AI is technically different and more challenging than for conventional systems. Establishing a cross-departmental review workflow and preparing test datasets requires enormous effort and time. As numerous AI applications and agent systems are introduced within companies, managing them all manually and maintaining audit trails is extremely difficult.
## Automating Governance Processes Based on Internal Roles
'Lens Governance' provides role-specific interfaces for application requesters (first line of defense, i.e., business units) and evaluators (second line, i.e., risk management and review departments). By controlling screen transitions according to each role's procedures, it enables a proper and swift AI governance review and approval process through collaboration.
Evaluators select appropriate datasets from the built-in 'Lens Safety Dataset' or the client's own datasets based on the application's risk level. The requester simply uploads the application's output based on this dataset, and the quality validation process proceeds automatically. The final approval is made by the evaluator, creating a 'Human in the Loop' configuration that balances automation with human oversight. The tool will also support verification based on legal systems and guidelines, such as Japan's AI Operator Guidelines.
Furthermore, the platform offers an inventory function (an 'AI asset ledger') to centrally manage and visualize which AI is running where and with what risk assessment. This feature also allows for scheduling periodic software asset audits and continuous quality assessments.
## Key Features of Lens Governance
- Automated approval processes and role-based screen transitions for evaluators (reviewers) and requesters (applicants).
- Inventory management to visualize AI utilization and risk distribution via a dashboard for management and evaluators.
- Scheduling for recurring and continuous validation, such as annual software asset audits.
- A built-in 'Lens Safety Dataset' covering 10 categories including safety, fairness, intellectual property, personal data leakage, and transparency for rapid and appropriate evaluation.
- Progressive support for datasets based on guidelines like the AI Operator Guidelines.
- Capability for custom evaluations based on client-specific metrics and use cases.
## About Citadel AI Inc.
Citadel AI is a global startup originating from Japan, focused on achieving the societal implementation of 'Trustworthy AI.' An international team, including engineers from Google's U.S. headquarters, leads development. The company provides proprietary AI governance tools that support the entire lifecycle of generative AI, from risk assessment at the implementation stage to guardrails during operation.