Conversational and Agentic AI for Identifying and Reducing Low-Value Care

Low-value care, including overuse, misuse and underuse of healthcare services, continues to place significant strain on health systems. Beyond clinical decision-making, many low-value interactions originate at the point where citizens seek information or reassurance, driven by uncertainty, unmanaged expectations and limited guidance. Health systems currently lack scalable, patient-facing tools that can intervene early and influence behaviour in a constructive and evidence-based way.

SESTEK positions the Virtual Agent as a frontline, agentic AI intervention operating at the interaction and behaviour layer of the healthcare system. By combining speech processing, LLM-based understanding and autonomous decision logic, the VA can filter unnecessary demand at first contact, explain when care is not required, promote high-value alternatives and proactively support preventive actions. Aggregated conversational data enables identification of recurring low-value patterns and access bottlenecks, supporting evidence-based decisions by providers and policymakers. The approach prioritises explanation and guidance over restriction, enabling sustainable behaviour change and more efficient use of healthcare resources.