HN Explorer

Show HN: OpenCareAI – AI Voice Agents for Healthcare (opencareai.com)

2 points by Achintya-dayal ↗ HN
I built a platform that lets healthcare providers create AI voice agents to handle patient calls and outreach. It integrates with phone systems via Twilio/Vonage and offers browser-based calling options. The agents can schedule appointments, answer routine questions, and follow up with patients. Everything is usage-based pricing, with full call logs and transcripts. Try it at: https://opencareai.com

4 comments

[ 3.3 ms ] story [ 24.9 ms ] thread
[dead] mattAHuman
[flagged]
Achintya-dayal
hey thanks for the comment. If you could elaborate more, would love to know your thought process.
mattAHuman
Is this a robot? If not imagine hot much you would enjoy hearing your test results from AI? “So doc, does this mean I have cancer”?? “Well it often does but it might not, blah blah blah,,” your AI is something patients will run from.

It is devoid of general intelligence and makes some of the most important conversations in life seem as if generative language replaces an MD, bedside manner, or a user interface. This may be an apt tool for the particularly UI challenged in a retail setting. I’ve implemented a few retail bots that can remind you of your recent orders for example, but even then the bot is basically a click path of one or two. Not a game changer unless your at SoftBank.

In a medical setting? It is the opposite of progress as it attempts to solve a problem that no one was asking for by applying an extremely misplaced tool;that is NLP/ generative language DNN. It’s regurgitated language as a masquerade for patient care and insight from decades of education and clinical experience. Oh and if the Bot is really a thin app around 3 options, guess what? Choosing the option is faster.

Achintya-dayal
Hey thanks for the detailed explanation. I agree with the fact that these technologies can never really replace human clinical intelligence,idea was to assist and not to replace. In India the doctor to patient ratio is at 1:1000. The idea is really simple "Can we give doctor more structured information about the patient before they walk in". The tech behind is more complex than just a simple NLP/generative language DNN.With a regular off the shelf nlp/generative language it is hard to create a clinical triaging questionnaire (technically a trained receptionist should do this work) but in India, things work differently. Just to set the context volumes of patients > trained staff in India. I completely agree that this can never replace actual patient care, but can it add value to the person who is providing it. Its being used for 3 different use cases currently, one of which is the tuberculosis patients in India are given reminder calls for taking their medicines on time - we've partnered with the TB society and resulting in 42% increase in medicine adherence. Let me know if you would want to use this tool for something else. Would be happy to give you a month's trial :). Your perspective was helpful, thank you.