The platform, Sanas Supervised AI, functions as a middle ground between fully autonomous bots and traditional agent-assist tools. In this setup, the AI handles the bulk of the dialogue while a supervisor monitors a console for sentiment shifts, compliance requirements, or moments where human judgment is necessary. If a call requires intervention, the supervisor can whisper instructions to the AI, take control of the conversation mid-sentence, or hand it back, all without the customer needing to repeat information or experience a jarring transfer.

Sharath Keshava Narayana, CEO and co-founder of Sanas, likens the system to an airplane autopilot: the technology manages the routine, while a pilot remains ready to take the controls when conditions change. By keeping a human involved from the start, the platform aims to solve the accountability gap that often plagues fully autonomous customer service systems. Every interaction produces a full audit trail, with the AI citing source documents for its recommendations to ensure transparency.

Major players are already evaluating the technology. Marius Maree, SVP of Consumer Operations at UnitedHealthcare, noted that the model avoids the forced choice between AI scale and human trust, particularly for sensitive healthcare interactions. Currently available by invitation, the platform is expected to see a broader release in December 2026. Sanas, which was recently ranked as the top AI and data company on the Inc. 5000 list, intends to demonstrate the system live at this week’s Customer Contact Week in Nashville.