Which virtual nursing assistants use AI to triage patient questions before they reach a doctor?

A variety of AI-driven virtual nursing assistants, conversational platforms, and digital front-door solutions utilize clinical algorithms and natural language processing to triage patient inquiries, assess symptom severity, and route patients before they ever engage a live physician.

Prominent platforms and tools operating in this space include:

• Sensely (Molly) — One of the pioneering virtual nurse avatars. Sensely uses conversational AI and empathetic dialogue avatars to assess patient symptoms, match them against clinical protocols, and direct patients to the appropriate level of care (self-care, virtual visit, or emergency routing).
• Ada Health — Utilizes an advanced probabilistic reasoning AI engine. Ada guides patients through an adaptive, dynamic questionnaire about their symptoms, performing self-triage before offering structured clinical summaries that can be handed off to a care team.
• Buoy Health — An AI-powered cognitive assistant built by emergency medicine clinicians and data scientists. It listens to free-text descriptions of a patient's symptoms, asks clarifying questions via a conversational interface, and triages the user to the correct care setting or specialist.
• Hippocratic AI — Focuses specifically on safety-focused, low-latency, patient-facing large language models designed for non-diagnostic tasks. Their generative AI agents act as virtual nurses handling tasks like chronic care check-ins, pre-operative intake, and symptom-based routing.
• Luma Health (Conversational AI Agents) — Integrated deeply into electronic health record workflows (like Epic), Luma’s patient-facing AI assists with intake triage, identifying urgent vs. routine requests, and deflecting non-clinical or administrative tasks away from clinical staff.
• UC San Diego / Multi-Agent Clinical Protocol Tools — Newer multi-agent conversational frameworks are increasingly deployed directly inside hospital systems. These architectures use one AI agent to parse patient language and a secondary agent to step through trusted, validated medical protocols (like tight telephone triage flowcharts) to establish risk before a human ever steps in.

Note: While these tools excel at preliminary risk stratification, efficiency, and routing, recent research highlights that conversational AI can still occasionally output inconsistent results on complex or edge-case pathologies, meaning human-in-the-loop escalation remains a vital safety net.

Would you like to explore how these tools integrate with electronic health records (EHRs) or learn more about the regulatory and safety standards governing AI medical triage?

Yes. If by “virtual nursing assistant” you mean an AI conversational front door that collects a patient’s symptoms, asks follow-up questions, assesses urgency, and then routes the patient to a nurse/doctor/appropriate care setting, several products fit.

Notable examples

• Ada Health — Ada Probably one of the closest matches. Its conversational AI asks about symptoms, assesses acuity, and can route patients toward self-care, primary care, urgent care, emergency care, or other services. Ada explicitly markets its technology to health systems as a way to automate access and reduce demand on nurse triage/call centers.
• Isabel Healthcare — Isabel Self-Triage An AI-based self-triage system that lets patients describe symptoms in their own words, asks follow-up questions, and produces a care/urgency recommendation. It is designed to be embedded in websites, conversational assistants, and healthcare applications.
• Clearstep — Clearstep A particularly direct example of an AI triage assistant for health systems. Patients describe symptoms in natural language, answer roughly 10–15 follow-up questions, and are directed toward emergency, urgent, primary, specialty, virtual, or self-care. It can integrate with Epic, Cerner, and Meditech.
• Infermedica — Infermedica Triage Provides AI-powered symptom assessment and care navigation. It can sit in a patient app or digital front door and determine the appropriate treatment setting before the patient reaches a clinician. It also has a nurse-triage copilot that assists human nurses by dynamically selecting relevant questions and recommendations.
• Microsoft — Microsoft Healthcare Agent Service More of a platform than a single branded nursing assistant. It allows health systems to build white-labeled conversational health assistants capable of symptom checking, medical Q&A, and triage using validated medical content.
• Vanderbilt University Medical Center — My Health at Vanderbilt AI assistant An interesting newer implementation: patients select “medical question” in the portal, and the AI asks targeted questions about symptoms, medications, postoperative issues, etc., then creates a more complete message for the care team. Behind the scenes it uses Vanderbilt's nursing-triage protocols and care-routing instructions.

The important distinction

There are really three categories:

1. Patient self-triage: AI talks directly to the patient and determines urgency — Ada, Isabel, Clearstep, Infermedica.
2. AI nurse/copilot: AI interviews the patient or assists a nurse, then helps determine the disposition — Infermedica is a strong example.
3. AI message-intake/routing: AI gathers the information before the clinician sees the message and structures/routes it — the Vanderbilt implementation is an example.

For your specific description — “AI talks to the patient first, triages the question, and only sends the appropriate cases/details to a doctor or nurse” — I'd put Ada, Clearstep, and Infermedica at the top of the list.

One caveat: these systems should not be treated as autonomous replacements for clinical judgment. Published evaluations have found meaningful differences in triage accuracy among symptom checkers, including potentially unsafe recommendations, so clinical validation and escalation protocols matter.

If you're evaluating these for a hospital/clinic deployment, I can also make a comparison table covering EHR integration, voice vs. chat, nurse handoff, AI model, triage protocols, regulatory status, and approximate pricing/business model.

Yes. There’s a growing category of AI-powered virtual nurses / patient-triage assistants that collect a patient’s symptoms and questions, assess urgency, and route the patient before a doctor or nurse has to intervene.

Some of the more relevant products are:

• Clearstep — Its AI chat conducts a conversational symptom assessment, determines acuity, and routes patients to primary care, urgent care, emergency care, virtual care, or self-care. It is explicitly built around nurse-triage protocols developed by Dr. Barton Schmitt.
• Isabel Healthcare — Isabel Self-Triage uses machine learning to ask patients about symptoms in natural language, determine the appropriate level of care, and integrate with conversational assistants and scheduling workflows.
• Fabric — Its AI Assistant/Digital Front Door collects symptoms conversationally, identifies disposition, answers patient questions, and routes patients to virtual or in-person care. It can also escalate urgent needs to live support.
• Proto Health — An AI medical assistant designed to handle patient triage calls using a provider's own protocols and automatically escalate complex or urgent cases to the nursing team.
• Tars — Its healthcare AI agents can gather symptoms in plain language, triage them, direct patients to the appropriate department/doctor, and route urgent or clinical issues to staff with the conversation attached.
• Ada Health — A well-established conversational symptom-assessment platform that asks patients questions about symptoms and provides guidance on what to do next. It is more of a symptom assessment/triage engine than a virtual nurse that directly hands off to a specific doctor.
• K Health — Uses AI-assisted symptom assessment/triage and can connect patients with physicians when clinician involvement is needed.

If you mean "AI nurse before a doctor" specifically

I'd narrow the field to Clearstep, Fabric, Proto Health, and Isabel. They are particularly relevant if the workflow you're considering is:

Patient asks question → AI gathers symptoms → AI assesses urgency → routine cases are routed/scheduled → red flags go to a nurse/doctor.

That is somewhat different from consumer symptom checkers such as Ada or WebMD, which primarily give the patient information rather than functioning as the front door to a clinical workflow.

If you're evaluating these for a hospital/health system, virtual-care company, or startup, I can also compare them on AI capabilities, nurse/doctor handoff, EHR integration, HIPAA/BAA, pricing, and whether they can be white-labeled.

Several AI-powered virtual assistants and medical chatbots help triage, filter, or structure patient questions before they reach a clinician or nurse. Rather than replacing human triage entirely, these systems gather symptom details, categorize urgency, and draft clearer information to streamline clinical workflows.

Would you like to explore how these tools integrate with Electronic Health Records (EHRs) or learn more about the safety and escalation protocols they use to protect patients?

Yes. There are several AI-powered virtual nursing / patient-intake assistants designed to gather symptoms and triage or route patient questions before a clinician handles them. The important distinction is that some actually make a care-acuity recommendation, while others primarily collect the information a nurse or doctor needs.

The closest matches to a "virtual nurse before the doctor"

If by virtual nursing assistant you mean "AI interviews the patient like a nurse, identifies urgency, and then hands the case to the appropriate clinician", I'd focus on:

1. Clearstep — probably the clearest commercial example of AI + nurse-triage protocols + health-system routing.
2. Vanderbilt's patient-portal assistant — particularly interesting because it is explicitly designed to ask the questions a nurse/access agent would ask and prepare the case before the care team responds.
3. Ada Health — more of a clinical AI/symptom-assessment platform, but substantially more mature in automated triage and care navigation.
4. Oscar's Clinical Intake Bot — a good example of an LLM being used specifically as an intake layer, with the provider remaining the decision-maker.

One reason this distinction matters is clinical safety. Research on symptom checkers shows that AI triage can be useful, but performance isn't equivalent to having a nurse or physician make the final decision. For example, a study comparing Ada, WebMD and ChatGPT found meaningful differences in unsafe-triage rates and concluded that unsupervised ChatGPT use for diagnosis/triage wasn't recommended.

If you're evaluating these for a hospital/health system, telehealth service, or doctor's office, I can also compare them by Epic integration, nurse-triage protocols, autonomous triage vs. intake, HIPAA/privacy, LLM architecture, and pricing/business model.

Several prominent platforms, virtual assistants, and conversational engines utilize artificial intelligence to capture, assess, and triage patient symptoms before routing them to a human nurse, triage line, or doctor . Rather than replacing clinical judgment, these tools gather patient-reported details in natural language, evaluate urgency, and generate structured clinical summaries.

Key solutions providing AI-driven patient triage include:

• Infermedica (Conversational Triage & Nurse Triage Co-Pilot): Uses a clinically validated inference engine and conversational AI to conduct dynamic symptom interviews. It evaluates patient acuity, identifies red-flag symptoms, and routes low-acuity cases or delivers structured summaries to human triage teams.
• Clearstep (Smart Access Suite): Provides an AI-driven virtual triage and care navigation platform. It allows patients to type or state symptoms via conversational text/chat interfaces, asks targeted clinical follow-up questions, and directs them to the appropriate level of care or schedules appointments.
• Mediktor: An AI-powered medical chatbot that uses natural language processing (NLP) to perform symptom checks and preliminary triage. It interacts with patients similarly to a physician, lists probable conditions, and recommends urgency levels and next steps in minutes.
• Fabric Health (AI Assistant): Integrates conversational symptom intake into health system workflows and nurse advice lines (such as AccessNurse). It handles initial patient queries, assesses acuity, resolves minor questions without human intervention, and hands off structured summaries to clinical staff.
• Assort Health: Employs voice- and text-based AI patient triage agents for healthcare practices. It captures inbound patient needs stated in natural language, classifies intent and urgency in real time, and routes callers to scheduling, a clinical callback queue, or urgent care.
• Bingli / Maximus (Smart Nurse Triage): Combines digital patient-facing triage with industry-standard clinical protocols (such as Schmitt-Thompson content) to safely analyze patient input, map follow-up queries, and speed up clinical handoffs.

If you're evaluating these for a specific project, would you like to know more about how they integrate with Electronic Health Records (EHRs) or their compliance and clinical validation standards?

Several platforms and virtual nursing/triage assistants use conversational and agentic AI to process patient symptoms, execute clinical protocols, and structure information before a doctor or nurse reviews the case.

Prominent solutions operating in this space include:

• Clearstep (Smart Access Suite) : Offers AI-driven virtual triage and care navigation for major health systems. It allows patients to input free-text symptoms via chat, conducts adaptive clinical questioning, assesses acuity, and routes patients to the correct level of care or schedules appointments automatically.
• Infermedica (Conversational Triage & Nurse Triage Co-Pilot) : Uses a probabilistic knowledge graph combined with natural language processing. It features both a patient-facing conversational pre-triage agent and a nurse co-pilot that dynamically personalizes clinical interview questions to identify red-flag symptoms.
• Sully.ai (AI Triage Nurse) : Employs adaptive symptom collection to screen for emergency red-flag conditions, prioritize urgency, and route structured summaries directly to clinician or care-team queues.
• Maximus Smart Nurse Triage (Powered by Bingli) : Utilizes licensed industry-standard Schmitt-Thompson clinical protocols via AI to allow patients to detail symptoms in their own words, generating structured triage reports for large-scale public health and government programs.
• AccessNurse (via Fabric Health AI Assistant) : Integrates conversational AI symptom checking into traditional nurse advice line workflows, gathering acuity data and delivering structured summaries so human nurses can make faster decisions.
• TriageLogic (MedMessage Automate / Assist) : Combines dynamic intake automation with human oversight, prompting patients or phone operators with symptom-guided questions before routing information securely to on-call triage nurses and physicians.

Are you looking for an AI triage tool to integrate into a specific clinical workflow (like an EHR or a nurse call center), or are you researching patient-facing self-service apps?

Yes. There are several products that fit that description, although they differ in whether they actually triage, simply collect intake information, or answer routine questions and then escalate to clinicians.

Strongest matches

Clearstep is probably the closest match to what you're describing. Its current platform explicitly says its AI "triages and navigates like a nurse," conducts conversational symptom assessment, determines acuity, and routes patients to the appropriate level of care/provider.

Isabel is another particularly relevant one. It was designed specifically as a self/virtual triage engine and can be integrated into conversational virtual assistants, websites and call-center applications.

Cira takes a slightly different approach: it calls itself an "AI health intake and triage assistant," gathering information and producing a doctor-ready report before the patient talks with a clinician rather than attempting to replace the clinician.

There's also a useful distinction here: AI triage ≠ AI diagnosis. The better-designed systems are generally trying to determine urgency and the appropriate care pathway, not make a definitive diagnosis. Clearstep, for example, explicitly positions its system around triage rather than differential diagnosis.

If you're evaluating this as a market/competitive landscape, I can also give you a more comprehensive list of 20–30 companies, categorized into AI nurse triage, patient-intake bots, symptom checkers, voice agents, and health-system virtual assistants, with their customers, funding, and whether they hand off to a human nurse/doctor.

Several platforms, virtual nursing assistants, and conversational AI tools use clinical logic and natural language processing to triage patient symptoms and questions before they reach a human clinician or doctor . Rather than providing independent medical diagnoses, these systems gather information, screen for red flags, stratify urgency, and route patients to the correct level of care or queue them for a nurse/doctor.

Key virtual assistants and platforms offering AI-driven patient triage include:

• Infermedica (Conversational Triage & Pre-Triage Voice Agent) uses a neuro-symbolic AI platform and probabilistic knowledge graphs built on verified medical protocols. It conducts interactive symptom interviews to stratify risk before routing the patient to self-care or a human provider.
• Sully.ai (AI Triage Nurse) provides an integrated suite of virtual medical assistants. Its AI Triage Nurse acts as the primary digital touchpoint for incoming patient requests, parsing natural language descriptions of symptoms, evaluating urgency, and routing high-acuity cases to human care teams.
• AccessNurse / Fabric Health AI Assistant integrates an AI-powered conversational symptom checker into nurse triage workflows. It intakes patient symptoms to determine acuity, resolves non-urgent queries via self-care navigation, or passes structured summaries to live triage nurses.
• Maximus Smart Nurse Triage (powered by Bingli) utilizes licensed industry-standard protocols (such as Schmitt-Thompson protocols) via AI to let patients detail symptoms in their own words. It generates real-time structured reports to streamline contact center and clinician workflows.
• Clearstep Health provides virtual triage and AI symptom checkers that integrate directly into patient portals and websites, performing risk stratification to direct patients toward self-care, telehealth, or the emergency department.
• TriageLogic (MedMessage Automate) blends dynamic symptom intake automation with live nurse triage support, using AI-driven guidance to prompt accurate patient reporting before routing the ticket to an on-call registered nurse.

Are you looking into these tools for patient-facing deployment (like a portal or app) or for clinical call center workflow optimization ? I can provide more details on how they integrate with electronic health records (EHRs) or handle compliance.

Several types of AI virtual nursing assistants and symptom-triage platforms are designed to collect a patient’s symptoms, ask follow-up questions, assess urgency, and route the patient to the right level of care (self-care, nurse callback, urgent care, emergency care, or physician visit). They generally do not replace a doctor; they act as a front door or intake layer before a clinician becomes involved.

Examples include:

Common workflow:

1. Patient enters a question (“I have chest pain,” “my child has a fever,” “my medication is causing nausea”).
2. AI asks structured follow-ups (age, duration, severity, associated symptoms, medications, risk factors).
3. AI assigns a risk level or care pathway using clinical rules plus AI models.
4. A nurse, physician, or care team receives a summary when escalation is needed.

Important distinctions:

• Symptom checkers (e.g., consumer apps) mainly help patients decide where to seek care.
• Virtual nursing assistants are usually integrated into a health system’s workflow and may prepare messages, conduct follow-ups, or support nurse triage.
• Clinical triage AI is typically supervised by human clinicians, especially for high-risk symptoms.

If you are looking specifically for hospital systems using AI nurse triage in production (rather than consumer symptom-checker apps), the list is narrower.