Perfecting Human-AI Interaction at Clinical Scale. Turning Production Signals into Safer, More Human Conversations
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915903095963648 |
|---|---|
| author | Mukherjee, Subhabrata Ausin, Markel Sanz Aggarwal, Kriti Datta, Debajyoti Puri, Shanil Jin, Woojeong Laud, Tanmay Manjunath, Neha Ding, Jiayuan Paudel, Bibek Schellenberger, Jan Huo, Zepeng Frazier Shen, Walter Shirazian, Nima Potter, Nate Perkari, Sathvik Filippova, Darya Morozov, Anton Mease, Austin Muppalla, Vivek Shakir, Ghada Miller, Alex Ghukasyan, Juliana Raglow-Defranco, Mariska Taylor, Maggie Mahal, Herprit Agnew, Jonathan |
| author_facet | Mukherjee, Subhabrata Ausin, Markel Sanz Aggarwal, Kriti Datta, Debajyoti Puri, Shanil Jin, Woojeong Laud, Tanmay Manjunath, Neha Ding, Jiayuan Paudel, Bibek Schellenberger, Jan Huo, Zepeng Frazier Shen, Walter Shirazian, Nima Potter, Nate Perkari, Sathvik Filippova, Darya Morozov, Anton Mease, Austin Muppalla, Vivek Shakir, Ghada Miller, Alex Ghukasyan, Juliana Raglow-Defranco, Mariska Taylor, Maggie Mahal, Herprit Agnew, Jonathan |
| contents | Healthcare conversational AI agents shouldn't be optimized only for clean benchmark accuracy in production-first regime; they must be optimized for the lived reality of patient conversations, where audio is imperfect, intent is indirect, language shifts mid-call, and compliance hinges on how guidance is delivered. We present a production-validated framework grounded in real-time signals from 115M+ live patient-AI interactions and clinician-led testing (7K+ licensed clinicians; 500K+ test calls). These in-the-wild cues -- paralinguistics, turn-taking dynamics, clarification triggers, escalation markers, multilingual continuity, and workflow confirmations -- reveal failure modes that curated data misses and provide actionable training and evaluation signals for safety and reliability.
We further show why healthcare-grade safety cannot rely on a single LLM: long-horizon dialogue and limited attention demand redundancy via governed orchestration, independent checks, and verification. Many apparent "reasoning" errors originate upstream, motivating vertical integration across contextual ASR, clarification/repair, ambient speech handling, and latency-aware model/hardware choices. Treating interaction intelligence (tone, pacing, empathy, clarification, turn-taking) as first-class safety variables, we drive measurable gains in safety, documentation, task completion, and equity in building the safest generative AI solution for autonomous patient-facing care. Deployed across more than 10 million real patient calls, Polaris attains a clinical safety score of 99.9%, while significantly improving patient experience with average patient rating of 8.95 and reducing ASR errors by 50% over enterprise ASR. These results establish real-world interaction intelligence as a critical -- and previously underexplored -- determinant of safety and reliability in patient-facing clinical AI systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_29893 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Perfecting Human-AI Interaction at Clinical Scale. Turning Production Signals into Safer, More Human Conversations Mukherjee, Subhabrata Ausin, Markel Sanz Aggarwal, Kriti Datta, Debajyoti Puri, Shanil Jin, Woojeong Laud, Tanmay Manjunath, Neha Ding, Jiayuan Paudel, Bibek Schellenberger, Jan Huo, Zepeng Frazier Shen, Walter Shirazian, Nima Potter, Nate Perkari, Sathvik Filippova, Darya Morozov, Anton Mease, Austin Muppalla, Vivek Shakir, Ghada Miller, Alex Ghukasyan, Juliana Raglow-Defranco, Mariska Taylor, Maggie Mahal, Herprit Agnew, Jonathan Human-Computer Interaction Artificial Intelligence Computation and Language Multiagent Systems Healthcare conversational AI agents shouldn't be optimized only for clean benchmark accuracy in production-first regime; they must be optimized for the lived reality of patient conversations, where audio is imperfect, intent is indirect, language shifts mid-call, and compliance hinges on how guidance is delivered. We present a production-validated framework grounded in real-time signals from 115M+ live patient-AI interactions and clinician-led testing (7K+ licensed clinicians; 500K+ test calls). These in-the-wild cues -- paralinguistics, turn-taking dynamics, clarification triggers, escalation markers, multilingual continuity, and workflow confirmations -- reveal failure modes that curated data misses and provide actionable training and evaluation signals for safety and reliability. We further show why healthcare-grade safety cannot rely on a single LLM: long-horizon dialogue and limited attention demand redundancy via governed orchestration, independent checks, and verification. Many apparent "reasoning" errors originate upstream, motivating vertical integration across contextual ASR, clarification/repair, ambient speech handling, and latency-aware model/hardware choices. Treating interaction intelligence (tone, pacing, empathy, clarification, turn-taking) as first-class safety variables, we drive measurable gains in safety, documentation, task completion, and equity in building the safest generative AI solution for autonomous patient-facing care. Deployed across more than 10 million real patient calls, Polaris attains a clinical safety score of 99.9%, while significantly improving patient experience with average patient rating of 8.95 and reducing ASR errors by 50% over enterprise ASR. These results establish real-world interaction intelligence as a critical -- and previously underexplored -- determinant of safety and reliability in patient-facing clinical AI systems. |
| title | Perfecting Human-AI Interaction at Clinical Scale. Turning Production Signals into Safer, More Human Conversations |
| topic | Human-Computer Interaction Artificial Intelligence Computation and Language Multiagent Systems |
| url | https://arxiv.org/abs/2603.29893 |