Supportiv’s Conversational Care® is powered by multi-patented technologies with nine proprietary AI models that go beyond understanding words to understanding concepts and emotions. Delivering proven care in ways no other system can achieve.
Recent advancements in chatbots and generative AI have put them in the spotlight, but they don’t and won’t ever truly support humans from an emotional health perspective.
Unlike rule-based systems, Supportiv’s understanding of mental, physical, emotional, and social health needs evolves, develops, and adapts continuously as new information is integrated through participant’s conversational interactions and moderator input.
Considering a participant’s emotional state, and creating a clear path to the exact right support and services at the exact right moment, is something only Supportiv’s technology has mastered. With the testimonials and outcomes to prove it.
Underpinned by a proprietary corpus of 178 million specialized data points from over 1.6 million users-to-date, Supportiv’s machine learning models are more precise than other systems. They are uniquely pristine and specifically fine-tuned for understanding the language of health and mental wellbeing — free of bias, trolling, sexism, racism and hate speech — thanks to rigorously-vetted multi-party feedback loops.
Supportiv’s AI precisely matches group chat participants within a vector of similarity based on their preferred conversation topic based on free-response language, providing the most effective and empathetic supporting experience.
Supportiv’s unparalleled content matching AI makes hyper-relevant resources instantaneously appear based on participants’ expressed needs using a pair of proprietary language models trained on over 170 million data points. The model improves over time using a feedback loop for error analysis, data labeling, and model retraining.
Our sophisticated, yet simple information architecture and design allows real-time conversation, plus a handy, non-distracting repository for resource and service referrals. All without anything getting lost or going out of focus.
Advancing upon existing measurements which track empathy delivered by an agent, Supportiv’s AI detects, tracks, and measures empathy *received* by each member of the peer-to-peer chats, to ensure each person is feeling the kindness, comfort, and support they need.
Using patterns found in thousands of manually-labeled messages, Supportiv uses a 110 million-parameter binary classification model to accurately predict which messages should be flagged for specialized action.