Natural Language Navigation™
What’s the first step when you seek medical care or mental health help?
Maybe it’s searching at random for something that might match what you vaguely think you need. Or, it’s: Please fill out these forms … clinical assessments … questionnaires … “quizzes”… Sigh!
Looking randomly, or answering those rote questions, are cumbersome and impersonal processes.
What if you could just say what’s going on in your life, simply talk about your pain, suffering, or worries, just describe it how you want to describe it, in your everyday speech –free from clinical terminology– and get the exact right kind of care you need, right in front of you, immediately?
Supportiv has developed the ability to bypass all forms and questionnaires and immediately route people to the most relevant service or resource, based on just a few words of expressed need, using its proprietary corpus of over 150 million natural language data points on the support people seek, and how they describe it, in their own words.
Supportiv partners with health plan and employer clients to fast-track members from their normal, everyday free text right to the most precisely relevant and helpful care or support with its Natural Language Navigation™ service.
How’s This Work?
By integrating Supportiv’s Natural Language Navigation™ service, you organization leverages Supportiv’s powerful machine learning models within your platform, portal, site or app, enabling your members’ written text or spoken words to return hyper-personalized resources, pages, or other materials from not just Supportiv’s own self-help library, but also your own full ecosystem of offerings.
For a new health plan or employer partnership, Supportiv:
- Crawls the list, directory, or pages of benefits, services, and/or resources available to members for pre-tagging and pre-processing.
- Leverages its proprietary language corpus to extrapolate all the myriad ways people might articulate need for each one of these possible service or resource matches, in their own regular, everyday, non-clinical language.
- Identifies a single question that will be used by the partner population as a prompt for the free text description of what they need help with. As examples, it could be “What’s your struggle?™”, “How can we support you?”, or “Tell us, what’s going on?”
- Leverages its patented capability to match a user’s unstructured text response to this question with the most relevant resource, in mere seconds, including presenting a user with the exact links or landing page for the right benefit, service, or resource.
Supportiv’s proprietary language corpus is the key. It’s:
- Free of noise, bias, trolling, hate speech, thanks to Supportiv moderators.
- Pure in its health and emotional wellbeing focus, specializing in the needs, moods, feelings, and emotions that arise from physical and emotional health challenges, as well as daily life struggles. .
- Able to robustly understand spelling errors.
With this proprietary corpus, Supportiv has trained multiple deep multi-layer neural networks, using both supervised and unsupervised machine learning techniques, to deliver Natural Language Navigation™ services for queries and resource sets of exceptional complexity.
Supportiv goes beyond the industry standard of unsupervised techniques such as word2vec, doc2vec, LDA or tf-idf. The unsupervised approach is suboptimal in that it does not allow for training the model on the specific task and data, nor continuous improvement of it via updating its training set on past mistakes.
Unlike the industry standard of processing search terms explicitly, Supportiv’s language models can understand the context, and even leverage back-and-forth conversation, for routing users to the most relevant service or resources.
Powering Forward
Natural Language Navigation™ from Supportiv avoids the pitfalls of traditional assessments (like engagement, bias, and validity issues), while creating streamlined, engaging access to care. This delivers better outcomes for both payers and members.
Less time and hassle is spent in the quest to access care. Once matched, people are empowered to help shape the kinds of recommendations others experiencing a similar struggle will receive, through iterative feedback.
Let Supportiv’s Natural Language Navigation™ technology curate the most appropriate services and most relevant resources for each member’s struggle–however it’s naturally described.
Contact us at [email protected] for more information on how your organization can leverage this technology.
