Proprietary hallucination-free AI agent architecture protects patients and providers in a proven enterprise-scale voice AI platform
Infinitus Systems, Inc., the AI company that has powered more than 100 million minutes of healthcare conversations that accelerate care for over a million patients, announced the first patient- and provider-facing voice AI agents built for use cases where trust is critical.
The Infinitus voice AI platform was built to ensure patients’ and providers’ trust in AI agents that can transform the patient experience with accurate, around-the-clock access to provider- or payor-curated expertise. Infinitus has advanced its proven multi-modal, multimodel platform for content-rich, often complex patient conversations that can include over 100 exchanges, ensuring they adhere to approved content and provider and plan procedures, as well as regulatory requirements.
Also Read: Epicrispr & Springbok Partner on AI in FSHD Human Study
Infinitus is the first voice AI architecture to include four innovations designed to build trust with providers and their patients:
- AI response control for trusted responses: A proprietary discrete action space leverages the power of LLMs to understand and facilitate conversations, while ensuring AI agents do not hallucinate or generate any responses that are inaccurate or outside of the approved standard operating procedure of customers or regulatory agencies.
- Specialized knowledge graph for trusted information exchange: The Infinitus knowledge graph allows agents to contextualize and verify information in real time, validating information provided by a patient or payor agent against trusted sources specific to the payor plan, treatment area, patient history, customer-provided knowledge bases, and more.
- AI review for trusted outcome: This automated post-processing and verification system incorporates human-level reasoning gathered from experience performing millions of tasks. The AI evaluates a conversation output to determine whether a response, as well as any extracted information, is accurate given the context of the task, providing a critical level of oversight that is impossible to achieve manually. When necessary, AI agents can flag a human and suggest follow-up actions.
- Robust security and compliance program for trusted processes: Infinitus adheres to SOC 2 and HIPAA requirements, which includes bias testing, protected health information (PHI) redaction, and data retention for secure, unbiased data handling.
While agents have been used widely in healthcare applications, they have not yet been purposefully designed to earn clinician and patient trust.
“As a patient, the most important questions come after hours, when care teams are off duty or asleep. AI agents can help reduce that anxiety and empower patients at any time, no matter where they are. But that can only happen when trust and accountability are built into the technology,” said Ankit Jain, CEO and co-founder of Infinitus. “We believe we are the first to deliver on the trust you should expect in healthcare at every level of our platform.”
Source: PRNewswire