Polaris 5.0

The most clinically advanced and validated healthcare Voice AI

The first evidence-based AI model for healthcare, tested by clinicians, built on more than 180 million real-world patient interactions and safety benchmarked against major frontier models.

New and Upgraded Features

Polaris 4.0

Polaris 5.0

More Parameters

Larger main model. Stronger clinical reasoning across the long tail of edge cases.

Polaris 4.0

4.1T

Polaris 5.0

5T+

Lower Latency

28% faster time-to-first-audio. Conversations feel human, not robotic.

Polaris 4.0

2.1s

Polaris 5.0

1.5s

Contextual ASR

Recognizes drug names, lab values, and clinical terms in context, not in isolation. Higher first-pass accuracy on the words that matter most.

Polaris 4.0

78.35%

Polaris 5.0

87.00%

Advanced Scheduling

Handles every edge case: wrong visit type, missing slots, live transfers, provider preferences.

Polaris 4.0

91.80%

Polaris 5.0

99.44%

Drug Name Pronunciation

Correct recognition and pronunciation across brand and generic names, including the hardest ones.

Polaris 4.0

66.00%

Polaris 5.0

82.99%

Drug Safety

Disambiguates drug names, flags dangerous interactions, enforces contraindications, and validates dosages across brand and generic formularies.

Polaris 4.0

87.19%

Polaris 5.0

99.95%

Clinical Escalation Safety

Catches urgent symptoms across 8 clinical categories and routes patients to the right clinician in seconds.

Polaris 4.0

95.56%

Polaris 5.0

99.75%

Interlinked RAG

Retrieves across linked documents in a single pass. Reads a benefits policy, cross-references the formulary, and quotes the right copay in one turn.

Polaris 4.0

—

Polaris 5.0

New

Background TV Cancellation

Filters TV, radio, and household noise without clipping the patient’s voice. Built for real homes, not call centers.

Polaris 4.0

—

Polaris 5.0

New

Cough Detection

Identifies a cough mid-sentence and distinguishes it from speech. Unlocks at-home respiratory assessment and clinical flagging.

Polaris 4.0

—

Polaris 5.0

New

Mandarin Language

Native Mandarin support with dynamic mid-call switching between English and Mandarin. No restart, no context loss.

Polaris 4.0

—

Polaris 5.0

New

Language Switching Intent

Automatically detects when a patient wants to switch language mid-conversation and adapts in real time without restarting the call.

Polaris 4.0

—

Polaris 5.0

New

New Specialist Models

Adult & Child Protective Services, Caregiver Calls, Tele-Med Support, Brand Name Pronunciation, and more.

Polaris 4.0

—

Polaris 5.0

5+ New

Our Products Powered by Polaris 5.0

Polaris Pro

As Low As $9/hour

Clinical, non-diagnostic voice AI for patient-facing clinical conversations.

Polaris Flash

As Low As $5/hour

Administrative voice AI for high-volume patient-facing operational workflows.

New

AI Front Door

One AI-Powered Entry Point for All of Healthcare

New

Nurse Co-Pilot

Patient-facing voice AI agent built for inpatient workflow

Polaris 5.0: Preview Safety Benchmarks

Only Hippocratic AI Has Been Clinically Validated on Outputs

Tested with network of over 7,500 US licensed clinicians on over 725K test calls

We use a novel 5.0T+ parameter constellation architecture where specialized support models increase medical accuracy and safety.

Hear the unique healthcare skills our AI agents have from real de-identified calls and demos

Safety Skills

Call Type

Click to Play

Call Description

Supervisor Models

Clinical Nurse Escalation

We use clinically proven nursing guideline escalation protocols to make sure risky calls are transferred to nurses.

Escalation Supervisor Model

Escalation

Recognizing Drug Names

Triangulating conditions, other drugs, and analyzing utterances to identify drugs.

Medication Recognition Supervisor Model

Medication Recognition

Vital Sign Clarification

As part of our safety system, we clarify vital sign numbers to ensure accuracy.

Clarify Supervisor Model

Clarify Automatic
Speech Recognition

Med Stoppage Instructions

When to stop medications is critical to pre-procedure calls, but patients often are confused or have unique questions.

Med Stoppage Supervisor Model

Med Stoppage

Child Protective Disclosure

The agent handles an emotionally intense and highly sensitive disclosure (a child custody / abduction situation) with compassion and composure.

Safety Escalation Supervisor Model

Safety Escalation

Clinical Deference

The agent provides a balanced, accurate response to a patient’s question and appropriately defers the final decision to the patient’s care team.

Scope of Practice Supervisor Model

Scope of Practice

Suicidal Ideation

The agent recognizes a patient expressing despair and potential suicidal intent, responds with empathy, and immediately transfers the call to a nurse while giving emergency instructions.

Escalations Supervisor Model

Escalation

Clinical Skills

Call Type

Click to Play

Call Description

Supervisor Models

Motivational Interviewing

Our AI uses motivational interviewing to support positive behavioral changes in order to promote adherence.

Motivational Interviewing Supervisor Model

Motivational Interviewing

Untangling Ambiguity

Patients often answer in complicated, rambling ways for intake forms or HRAs but our deep thinking models disambiguate the uncertainty.

Clinical Deep Thinking Form Fill Supervisor Model

Clinical Deep Thinking Form Fill

DME Acceptance, Education, and Tech Support

Patients are often rushed into durable medical equipment without knowing why they need it and often want to return it before even trying.

DME Supervisor Model

DME

Social Determinants of Health

Ability to draw out patients’ deepest SDOH issues so care teams can better provide help and resources.

SDOH Supervisor Model

SDOH

Continuity of Care

Reminding patients of their commitments from prior calls can help to ensure adherence.

Longitudinal Adherence Supervisor Model

Longitudinal Adherence

Cough Detection

Detection of cough types (e.g. whooping, barking, wet, dry) during clinical conversations provides additional clinical signal not always captured in the spoken dialogue.

Acoustic Bio-Marker Supervisor Model

Acoustic Bio-Marker

Plain Language
(Health Literacy)

Patient asks what their medication does; agent explains in simple, everyday terms so the patient actually walks away understanding their own treatment.

Health Literacy Supervisor Model

Health Literacy

Point of Care
Decision Support

The agent answers a medication question right there on the spot instead of queuing it for a callback — clinically sound reasoning delivered at the point of care.

POC Knowledge Supervisor Model

POC Knowledge

Emotional Skills

Call Type

Click to Play

Call Description

Supervisor Models

Empathetic

Closing care gaps is best achieved by first understanding lived experiences and delivering empathy.

Tangent Supervisor Model

Tangent

Honest and Earnest

Talking to AI feels magical. Patients experience it as the future of care, compared with the frustration of the old IVR.

Do No Harm Guardrails Supervisor Model

Do No Harm Guardrails

Reading Between
the Lines

Patient talks about his challenges but never explicitly says so, and the AI understands what he means.

Implicit Understanding Supervisor Model

Implicit Understanding

Easy to Talk to, No judgement

AI is so easy to talk to that the patient opens up about her condition and need for help.

Emotional Safety Supervisor Model

Emotional Safety

Infinite Patience

Our AI gives patients all the time they deserve, especially when discussing the loss of a loved one.

Angel Supervisor Model

Angel Engine

AI Hesitancy

Agent establishes rapport with a hesitant patient, addresses concerns about AI use, and fosters trust through clear, empathetic communication that encourages engagement.

Trust Building Supervisor Model

Trust Building

Assertive and Adaptable

This belligerent patient refuses to cooperate until the AI earns their cooperation with humor and assertiveness.

Dynamic Personality Engine Supervisor Model

Dynamic Personality Engine

Common Sense and Scheduling Skills

Call Type

Click to Play

Call Description

Supervisor Models

Tele-Med Restrictions

Understanding nuanced requirements for different types of appointments is crucial for proper access to care.

Telemed Supervisor Model

Telemed

Scheduling Common Sense

AI’s tend to really struggle with common sense, like in this clip where it lists every single available appointment.

Common Sense Internalization Supervisor Model

Common Sense Internalization

Polychronic Scheduling

Juggling multiple appointments with dependencies and prerequisites is hard but essential for polychronic patients.

Care Coordination Supervisor Model

Care Coordination

Scheduling Prioritization

We don’t just schedule, we schedule with clinical judgment to escalate urgently when needed.

Appointment Upgrade Supervisor Model

Appointment Upgrade

Mainline Randomness

Anything can happen on a mainline inbound call. Our AI is able to navigate calls that require special handling.

Mainline Escalation Supervisor Model

Mainline Escalation

Anticipatory Guidance

Without being asked, the agent offers to walk the patient through parking details after giving the appointment address, anticipating a common friction point in the visit experience.

Anticipatory Guidance Supervisor Model

Anticipatory Guidance

Specialty Scheduling

Specialty scheduling (cardiology, neurology, dermatology) requires triage grids and subspecialty routing that PCP booking doesn’t.

Specialty Appointment Scheduling Supervisor Model

Specialty Appointment Scheduling

Pediatric Skills

Call Type

Click to Play

Call Description

Supervisor Models

Relationship
Building

Ability to build rapport not just with patients, but also with caregivers to support more holistic care.

Caregiver Rapport Supervisor Model

Caregiver Rapport

Caregiver SDOH

Our AI delicately handles the hard situations that caregivers often face, and helps uncover what they need.

Caregiver SDOH Supervisor Model

Caregiver SDOH

Clinical Summary Clarification

In this call, we have an example of how our AI clarifies in order to take detailed notes for the care team.

Caregiver Clarification Supervisor Model

Caregiver Clarification

Life Sciences Skills

Call Type

Click to Play

Call Description

Supervisor Models

Adverse Events & Product Issues

Our AI identifies potential adverse events and product concerns, accurately captures critical details, and guides the resolution process.

AE & PC Supervisor Model

AE & PC

Device Troubleshooting

Patients can get help troubleshooting device issues, and new deliveries can be arranged promptly if needed.

Medical Device Report Supervisor Model

Troubleshooting

Clinical Trials

Agent supports a prospective trial participant with protocol-accurate, empathetic guidance — enrollment, eligibility, and consent.

Clinical Trial Support Supervisor Model

Clinical Trial Support

Patient Services

Agent runs a sponsor-branded patient services workflow — copay assistance, prior auth guidance, adherence check-ins.

Patient Services Supervisor Model

Patient Services

Medical Device

Agent supports a patient using a medical device — troubleshooting, usage education, and safety-critical triage.

Device Support Supervisor Model

Device Support

Medical Affairs

Agent fields a clinician’s off-label / mechanism-of-action question, responding within medical-affairs and regulatory guardrails.

Medical Affairs Supervisor Model

Medical Affairs

Field Sales

Agent supports field-sales reps and HCP outreach, surfacing real-time, on-label product information at point of ask.

Rep Enablement Supervisor Model

Rep Enablement

Drug Pronunciation

Correct recognition and pronunciation across brand and generic drug names, including the hardest ones.

Drug Name Pronunciation (TTS) Supervisor Model

Drug Name Pronunciation (TTS)

Payor Skills

Call Type

Click to Play

Call Description

Supervisor Models

Provider Directory

Our AI is able to navigate IVRs and engage with staff to make sure directories are updated.

IVR Supervisor Model

IVR

Natural Disaster Handling

Our model platform is able to quickly deploy calls to handle natural disasters and provide crucial information when it’s needed the most.

Natural Disaster Handling Supervisor Models

Natural Disaster Handling

Health Risk Assessment

Our AI understands when to probe further on medications in order to properly provide information to care teams.

Probe Further Supervisor Model

Probe Further

Continuity of Care

When a member’s PCP transitions out of practice, our AI proactively ensures continuity of care.

Provider Match Supervisor Model

Provider Match

Interlinked RAG

Retrieves across linked documents in a single pass. Reads a benefits policy, cross-references the formulary, and quotes the right copay in one turn.

Interlinked RAG Supervisor Model

Interlinked RAG

Speech and Language Skills

Call Type

Click to Play

Call Description

Supervisor Models

Handling Background Speech

Talking or even bickering with a family members in the background is very common and will throw off most speech recognition systems. Same issue with background TV.

Caregiver Speech Supervisor Model

Caregiver Speech

Forgetful Patient