Scalable & Robust
Agent Management
Address Technical Challenges
Conduct of collaborative research to develop, validate, and automate methods for agentic AI evaluation, monitoring, and remote oversight
Bring together key stakeholders from industry, academia, and government to foster the adoption of scalable AI Management Systems (AIMS)
Support organizations operating in high-risk environments in their journey to streamline, strengthen, and demonstrate Agentic AI governance
Engage Stakeholders
Automate Responsibility
ConverGen AI enables the responsible deployment of clinical agents through research and advisory services
Research Collaborations
We create value through productive collaborations with healthcare and technology providers who are solving the known and emerging challenges in developing, evaluating, and deploying Agentic AI solutions in healthcare
Evaluating the role of AI Management System (AIMS) certification in healthcare
Comprehensive assessment of the role that ISO 42001-Certified AI Management Systems (AIMS) have in the responsible deployment of clinical GenAI solutions
Understanding why clinical GenAI solutions require a form of oversight that is distinct from Narrow AI's
Non-technical description of the fundamentally distinct nature of GenAI, which drives the need for an alternative regulatory oversight approach centered on AIMS certification
Ongoing Projects
Clinical Evaluation of AI Agents
Automated Agent Management
Optimized Inputs to Agentic Workflows
There is a major need to ensure the ongoing safe operation of Agentic solutions. The main challenge is that we can’t evaluate them solely as human counterparts or as regular static products. The best approach to evaluate Agents’ safety and performance should combine new, purpose-built technical metrics with automated, continuous assessments that focus on clinical and operational outcomes.
ConverGen AI conducts collaborative research to develop and validate methods for effectively evaluating Agents in real-world clinical settings.
The idea of sourcing, implementing, and managing Agents is daunting for most organizations, as they lack clear guidance in a rapidly evolving field with considerable uncertainty. It is also clear that they can’t simply have “trusted” humans overseeing all Agentic solutions, as that would be prohibitively expensive, hard to scale, and ultimately ineffective.
ConverGen AI develops methods to streamline the processes used by organizations to envision, set up, monitor, and continuously improve agentic solutions through auditable ISO/IEC 42001-compliant automated AI Management Systems (AIMS).
As agents assume core business operations in high-risk settings, they will naturally become the target of external oversight. While privacy mandates should prevent complete transparency, it is still feasible and desirable to enable external oversight through abstracted metrics related to safety, effectiveness, and compliance with the ISO/IEC 42001 standard for Agentic AI Management Systems (AIMS).
ConverGen AI develops workflows and methods to enable streamlined continuous external oversight of mission-critical agentic functions in high-risk industries.
The implementation of safe and effective Agents mainly depends on the quality of operational data inputs and the strength of the workflow. The GenAI models have an important, but secondary, role as they operate based on the information, business context, and guidance provided to them. Poor or misleading data, as well as careless or broken workflows, inevitably result in waste and disappointment.
ConverGen AI develops methods to ensure Agents act based on high-quality, disambiguated data and adhere to principles carefully established by the organizations legally responsible for their actions.
