Citadel AI
Citadel AI provides automated testing and monitoring software to evaluate the quality, safety, and compliance of various AI systems, including traditional machine learning models and advanced Generative AI systems and Agentic AI applications used in healthcare. Our tool, Citadel Lens, enables rigorous technical and compliance testing against international standards and regulations (e.g., ISO/IEC, EU AI Act). As a CHAI-certified AI Assurance Resources Provider, Citadel AI helps medical institutions, technology vendors, insurers, and auditors ensure AI systems are safe, fair, and trustworthy.
Citadel AI provides automated testing and monitoring software to evaluate the quality, safety, and compliance of various AI systems, including traditional machine learning models and advanced Generative AI systems and Agentic AI applications used in healthcare. Our tool, Citadel Lens, enables rigorous technical and compliance testing against international standards and regulations (e.g., ISO/IEC, EU AI Act). As a CHAI-certified AI Assurance Resources Provider, Citadel AI helps medical institutions, technology vendors, insurers, and auditors ensure AI systems are safe, fair, and trustworthy.
For model deployers/users (e.g. health systems): Citadel AI helps medical institutions validate vendor-supplied or internally developed AI tools before deployment, monitor them continuously in production, and generate evaluation reports to support governance and procurement decisions.
For model developers (e.g. med-tech): Citadel AI enables developers to test models and identify gaps in performance and compliance, and provide audit-ready reports for customers and regulators.
For insurers (e.g. payors, insurance providers): Citadel AI delivers independent validation evidence that supports AI-based decisions, reduces liability risk, and ensures that AI-enabled care solutions meet safety and quality requirements.
For auditors/regulators (e.g. certification bodies, auditing firms): Citadel AI tests AI systems against international standards and produces standardized audit-ready reports that make it easier to review AI systems.
In-scope
Citadel AI’s assurance resources are intended for testing and monitoring the quality, safety, and compliance of AI systems used in healthcare. This includes:
Evaluation of internally developed models by medical institutions
Validation of vendor-supplied AI systems prior to procurement and deployment
Generation of standardized reports to support governance, regulatory submissions, and reimbursement decisions
Out-of-scope
Citadel AI does not develop or sell clinical AI models. We also do not act as a data custodian or infrastructure provider for health data. All testing is performed with customer-provided datasets in SaaS or customer-secured environments.
Limitations
Citadel AI’s assurance resources are software-driven; while we support project-based testing sprints, our primary business model is a 12-month software license for continuous monitoring.
Testing outcomes depend on the quality and representativeness of data provided by the customer.
We do not certify regulatory approval; instead, we provide transparent evaluation and reporting that can be used as supporting evidence for regulatory and governance processes.
Citadel AI supported 50+ customers and projects globally as of September 2025. Citadel AI works not only with AI users such as health systems but AI auditors such as British Standards Institution (BSI), Deloitte, and PwC to ensure AI compliance.
Traditional AI
Machine Learning (ML)
Generative AI
Agentic AI
Citadel AI primarily operates a seat-based 12-month software license model, which provides ongoing monitoring capabilities. For CHAI members, we also offer short-term “AI Testing Sprints” at a project-based fee, subject to individual price quotation, designed as a proof-of-concept. Medical institutions and technology vendors can then seamlessly transition to an annual license for continuous assurance.
Citadel AI operates as an independent assurance provider and does not develop or sell clinical AI models. We do not hold ownership stakes in our customers’ AI products. Our role is strictly to provide objective testing, monitoring, and assurance services in alignment with CHAI’s and international standards.
Citadel AI provides software tools to enable organization-wide AI governance and does not provide datasets. Citadel AI tests datasets developed and prepared by user organizations. We can also collaborate with 3rd-party test dataset providers.
Citadel AI is SOC2 equivalent ISO27001 certified.
Citadel AI works globally with healthcare providers, med-tech companies, insurers, and regulators to strengthen the trustworthiness of AI. We actively contribute to international initiatives on AI quality and compliance. Our mission is to make safe and reliable AI a foundation of modern healthcare.
