Eleos Launches Polaris AI, Revolutionizing Post-Acute Care with AI-Powered Conversations

Strategic Insights -

Boston-based Eleos, the leader in AI for post-acute care, has announced the launch of Polaris AI, a cutting-edge AI model built in collaboration with Google Cloud. Polaris AI is designed to navigate the complexities of post-acute care conversations, recognizing evidence-based techniques, handling sensitive disclosures, and managing long, complex sessions. This advanced model is audio-native, capable of interpreting vocal cues and recognizing nuance, emotional tone, and implicit meaning in authentic provider-client conversations.

Key Takeaways:

  • Polaris AI is built on Google Cloud's secure, scalable infrastructure and is enabled by its family of multimodal models, combining advanced reasoning and multimodal capabilities with deep domain expertise in post-acute care.
  • The model is trained on a diverse, clinically de-identified corpus of millions of minutes of therapy transcripts, behavioral health progress notes, professionally tagged datasets, and proprietary in-house content to equip clinicians with a tool that focuses on providing care.
  • Polaris AI can recognize evidence-based techniques such as CBT, DBT, MI, and ACT within sessions, as well as handle and flag sensitive disclosures without hallucination or redaction of critical details.
  • The model is optimized to understand real-world clinical dialogue, recognize nuance, emotional tone, and implicit meaning in authentic provider-client conversations.
  • Security is a top priority, with Polaris AI being deployed on Google Cloud's trusted infrastructure that supports HIPAA compliance and has built-in AI defenses and continuous monitoring to protect patient and provider data.
  • Protections include continuous AI stack monitoring, secure development lifecycle practices, identity and access management, and no storage of raw audio or transcripts, paired with a full-time CISO and dedicated security team.
  • Polaris AI reflects Eleos' belief that technology should honor the human side of care and marks the next step in its sustained, multi-phase effort to set a new benchmark for accuracy, compliance, and efficiency in post-acute care environments.

Statistics:

  • Polaris AI is trained on a diverse, clinically de-identified corpus of millions of minutes of therapy transcripts and behavioral health progress notes.
  • The model can recognize evidence-based techniques such as CBT, DBT, MI, and ACT within sessions.
  • Polaris AI can handle and flag sensitive disclosures without hallucination or redaction of critical details.
  • The model reduces documentation time by more than 70%.
  • Client engagement is boosted by 2x, and symptom reduction is improved by 3-4x.

Sources:

  • GlobeNewswire: "Eleos, Leader in AI for Post-Acute Care, Announces Launch of Polaris AI" (October 20, 2025)
  • Eleos: "Polaris AI: Revolutionizing Post-Acute Care with AI-Powered Conversations"
  • Google Cloud: "Polaris AI: Built on Google Cloud's Secure, Scalable Infrastructure"
  • Centerstone: "Prasad Kodali, Chief Information Officer, on Eleos' Security Infrastructure"
  • Eleos: "About Us: Eleos is the Leading AI Platform for Behavioral Health, Substance Use Disorder, Home Health, and Hospice"