Clearstep

What is Clearstep?

Clearstep is a clinical AI and patient navigation platform designed for hospitals, medical systems, and patient access centers. It utilizes chat, voice, and workflow agents to carry out tasks such as symptom collection, triage, determination of the appropriate treatment route, scheduling appointments, insurance verification, and coordination of referrals.

The product is operated by Clearstep Inc., a company founded in 2018 with its headquarters in Chicago, United States. It is primarily sold as infrastructure for digital front doors and contact centers in medical institutions, rather than as a health chat application that ordinary consumers can subscribe to on their own.

Important medical advice

The outputs of Clearstep are intended for educational purposes and as a tool for decision support; they cannot replace a doctor’s diagnosis, treatment recommendations, physical examination, or a comprehensive clinical assessment. The official terms state that natural language systems and symptom checkers may make mistakes, and no guarantee is given regarding the accuracy and reliability of their results.

If a medical emergency is suspected, contact the local emergency services or qualified medical professionals immediately; do not wait any longer for AI-based responses. Doctors should also not rely on content from such platforms in place of their own professional judgment.

A one-sentence summary

Clearstep makes use of clinical AI based on the Schmitt nurse triage protocol; it allows patients to describe their symptoms in plain language and answer relevant questions, after which it directs them to the appropriate service – whether it is emergency care, urgent care, primary care, specialist care, telemedicine, or home care.

Main functions

AI Chat and symptom triage

The patient first describes their symptoms in free text, after which the system asks around 10 to 15 clinical-related questions. According to the information on the product page, the entire process is completed within a few minutes, and the appropriate level of medical care as well as the next steps are suggested.

• Identify the main symptoms, duration, severity, and accompanying manifestations.
• Adjust subsequent questions based on age, gender, medical history, and responses.
• Screening is needed to identify dangerous signs that require emergency care or immediate contact with a professional.
• Redirect non-urgent cases to primary care, specialized services, telemedicine, or self-care.
• Actual navigation is carried out by taking into account the services available in the medical system, the location, and the resources that can be booked.

Schmitt clinical protocol

Clearstep triage is based on Dr. Barton Schmitt’s telephone triage protocol, which has been in use for many years at the American Nurses Consulting Center. This protocol provides a structured clinical approach, but digital implementation and patient descriptions still cannot eliminate errors, misinterpretations, or cases that are difficult to classify.

AI Voice voice triage

AI Voice applies the same clinical triage logic to telephone channels, and can be used for nurse consultation hotlines as well as medical contact centers. It listens to the symptoms described, asks follow-up questions, determines the appropriate course of action, and transfers the call to a human operator when necessary.

Care Navigation Patient Navigation

Care Navigation helps patients find the right doctors, locations, services, and appointments, rather than merely providing an abstract rating of medical facilities. Medical systems can incorporate data on their facilities, networks, services, and capacity into this routing process.

Appointment and insurance verification

Smart Access allows the results of the triage process to be linked to scheduling arrangements, as well as checks regarding eligibility and insurance, thereby reducing the need for patients to repeat their information across various pages or over the phone. The actual range of available appointments depends on the healthcare institution’s systems and available resources at any given time.

AI Intake and clinical decision support

Clearstep can collect patient information, symptoms, and structured responses prior to a consultation, providing decision-support information for the clinical team. The final diagnosis, medical orders, and treatment remain the responsibility of qualified medical professionals.

Contact center automation

The platform is capable of handling a high volume of phone calls or chat requests, carrying out issue identification, triage, and routing tasks in advance for human operators. For complex situations, those involving emotional distress, unclear communication, or high risk, there should be an option for quick escalation to human intervention.

Capacity optimization

The Capacity Optimization Suite combines patient needs with factors such as outpatient services, locations, doctors, and appointment capacities in order to predict demand, carry out dynamic scheduling, and achieve load balancing across different sites. Its purpose is to reduce bottlenecks and unused resources, rather than changing clinical priorities.

Multi-channel and white-label experiences

Healthcare institutions can integrate Clearstep into their websites, patient portals, mobile applications, or customer service centers, using their own brand. The channels for access, the process of logging in, identity verification, and the level of accessibility are determined by the specific implementation plan.

Patient triage workflow

1. Patients access the Clearstep process through the healthcare institution’s website, apps, chat, or phone.
2. The system prompts the use of boundaries and collects information on age, symptoms, and relevant background details.
3. The patient describes the main issues in natural language, and the AI identifies the relevant clinical concepts.
4. The system proceeds to ask 10 to 15 targeted questions based on the previous answer.
5. Clinical protocols are used to identify risk signals, determine the urgency of seeking medical care, and decide on the appropriate level of service.
6. The platform offers actionable options by taking into account the location of the facilities, doctors, the network infrastructure, and capacity.
7. Patients can continue to schedule appointments, switch to a human agent, contact a nurse, or seek emergency assistance as indicated.
8. The triage summary is entered into the relevant systems according to the configuration of the institution, for use by subsequent teams.

Tutorials for medical institutions

Launch the pilot project

1. Choose a clearly defined entry point, such as diversion for non-urgent adult symptoms or a nurse hotline.
2. Form a joint team comprising clinicians, nurses, patient visit specialists, privacy officers, security experts, legal professionals, and IT staff.
3. Define rules for emergency escalation, manual takeover, language, age, and symptoms that are not applicable.
4. Organize information on agency services, locations, insurance networks, booking templates, and real-time capacity data.
5. Connect EMR, scheduling systems, and CRM in the testing environment, and use masked patient cases for verification.
6. Have clinical staff conduct blind reviews of the triage results, and record cases of over-triage and under-triage.
7. Launch on a small scale and monitor completion rates, upgrade rates, deviations, complaints, and security incidents.
8. Gradually expand the target audience and distribution channels once the safety thresholds set by the organization are met.

Verify the voice AI process

1. Simulated calls using different accents, speech speeds, background noise, and phone quality.
2. Test negative expressions, drug names, complex medical histories, and sudden changes in symptoms.
3. It ensures that key questions are not skipped in cases of repetition, silence, interruption, or failed recognition.
4. Identified danger signals can be immediately upgraded to manual or emergency alerts.
5. Have the clinical team rate transcription, triage, and routing separately.
6. Set rules for maximum waiting time, retry on failure, manual intervention, and call history.

Connect to hospital systems

1. Define the responsibility boundaries for each system, including patient identification, medical records, scheduling, insurance, and CRM.
2. Use service accounts with the minimum required permissions, and separate the permissions for reading, scheduling, and writing medical records.
3. First, read the information of the patients undergoing testing and their test numbers; do not proceed with actual appointments directly.
4. Verify abnormal situations such as duplicate patients, name changes, expired insurance, and time zones.
5. Field mapping, channel identification, and manual correction mechanisms are provided for the write-back summary.
6. Go live only after completing the security review, audit logging, disaster recovery, and downtime procedures.

Which institutions are suitable?

• Large hospitals and medical systems: Establish a unified digital front door as well as a system for routing patients between different hospital locations.
• Nurse consultation hotline: First, complete the collection of standard questions, and then refer high-risk or complex cases to nurses.
• Medical Contact Center: Automatically handles triage, navigation, scheduling, and common administrative requests.
• Primary care and specialty networks: directing patients to the appropriate doctors, locations, and types of appointments.
• Health plans and payers: assist members in finding services within the network as well as the appropriate level of medical care.
• Telemedicine platform: Collects symptoms and determines suitability prior to the video consultation.
• Medical services for the general public or specific groups: 24/7 chat and voice support available.

Typical use cases

• The patient is unsure whether to go to the emergency room, urgent care, or make an appointment with a family doctor.
• There is not enough staff available to answer inquiries about common symptoms at night or on weekends.
• The contact center needs to first determine the clinical urgency before deciding on the queue.
• Patients need to find services within the network that are nearby and have available slots.
• The hospital wishes to collect structured information on symptoms and medical history before making an appointment.
• A multi-location system needs to balance appointments and patient arrivals based on real-time capacity.
• Spanish-speaking patients need to use a localized self-service triage experience.
• The nursing team wishes to separate repetitive administrative tasks from high-risk clinical consultations.

Product advantages

• It is designed around patient interactions in medical systems, rather than a generic chatbot framework.
• It is based on established nurse triage protocols, and combines clinical issues with the actual pathways for seeking medical care.
• The triage logic for chat and voice communication applies to both digital channels and traditional phone calls.
• It not only suggests the appropriate level of medical care, but also provides access to information on doctors, locations, appointment scheduling, and insurance options.
• It can be integrated with medical systems such as Epic, Oracle Health Cerner, and MEDITECH.
• It supports the customization of institutional brands, service networks, and capacity rules.
• Integrate clinical triage, patient navigation, and operational capacity on the same platform.
• The official website’s security page provides information on encryption, penetration testing, and cloud infrastructure practices.

Clinical validation and accuracy

The current homepage of Clearstep claims a clinical accuracy rate of over 97% as well as the ability to complete triage within 3 minutes; earlier official materials, however, cited an accuracy rate of over 95%. These different figures may stem from various versions, sets of cases, or assessment methods, and therefore cannot be regarded as a guarantee for all actual patients.

The symptom assessment page indicates that patients typically answer 10 to 15 questions, while the Smart Access page states that the entire process takes less than 5 minutes. Medical institutions should request to see the full research design, the composition of the patient cases, the criteria used for comparison, the confidence intervals, and the performance levels for each risk category.

• The overall accuracy rate cannot replace a separate assessment of inadequate triage and serious missed diagnoses.
• The test set needs to include individuals from child, elderly, pregnant and nursing populations, as well as those with chronic diseases and complex comorbidities.
• The English, Spanish, voice, and text channels should be evaluated separately.
• Real-world monitoring should focus on different genders, races, regions, and groups with varying access levels.
• Agreement on the protocol does not necessarily mean an improvement in clinical outcomes; it is still necessary to monitor the referral and subsequent results.

Usage restrictions and precautions

• Clearstep does not provide medical advice, diagnosis, or treatment, nor can it replace a doctor.
• Natural language models are unable to understand all expressions correctly; omissions or misstatements by patients can affect the results.
• The terms of service are primarily intended for residents of the United States; availability and legal compliance in other regions are not guaranteed.
• The product is mainly deployed through medical institutions; there are no separate purchase packages available for individual use.
• In case of critical symptoms, it is necessary to contact the local emergency services immediately; there should be no reliance on continuing with question-and-answer sessions.
• Accuracy is specific to certain validation results and does not apply to every symptom, population group, or channel.
• The current focus of Spanish language experiences is on Mexican Spanish; other variants require verification.
• Age, gender, medical history, and symptoms are considered sensitive health data, and the principle of minimization should be followed.
• Connecting the system to EMR and scheduling appointments can have real-world consequences, so strict access controls and auditing are necessary.
• The public page does not provide any FDA licensing or approval information that can be verified directly; therefore, such claims should not be made.
• The public security page does not list all compliance certifications; the status regarding HIPAA and BAA should be determined based on the procurement contract.
• The old privacy policy stated that aggregated, anonymized health data could be used to improve algorithms; organizations should therefore review their current data protocols.

Price and procurement methods

Clearstep does not disclose the prices for individual subscriptions or standard corporate packages; its website mainly directs medical institutions to schedule demonstrations. Implementation involves various elements such as modules, distribution channels, the number of patients, integration, clinical configurations, security reviews, and support, which is why customized quotes are provided.

The purchaser should request that separate quotes be provided for licensing, implementation, interfaces, SMS or voice services, cloud hosting, support, upgrades, and the maintenance of clinical content. It is also necessary to clarify the minimum contract duration, requirements in case of excess usage, data export procedures, responsibilities regarding the termination of services, and model updates.

Support channels and integration

APIs, SDKs, and open-source status

Clearstep’s public examples show that digital triage can be integrated into medical systems via APIs, and they also mention integration with EMR, scheduling, facility management, and CRM databases. However, no public endpoint documentation, self-service API keys, pricing information, or developer SDKs are available; it seems that these interfaces are intended for use within corporate environments.

No verifiable official GitHub organization, product source code, or open-source license was found. The Clearstep platform, the clinical logic components, and the AI modules should be considered non-open source; third-party medical open-source projects cannot be regarded as part of its official code.

Privacy and data security

Clearstep collects symptoms, age, gender, medical history, and other health information provided by patients voluntarily; it also records the IP address, browser type, postal code, search terms, and time of interaction. This health data is used to provide more relevant information and better advice regarding medical care, and it may be analyzed in an aggregated and anonymized form to understand user behavior and improve matching algorithms.

The privacy policy states that certain health information can be linked to personal data in order to avoid repeated entry. The data is governed by U.S. laws, and when users outside the United States submit information, it may be transferred to the U.S.

• All customer data is encrypted both at rest and during transmission.
• The platform uses AWS KMS and hardware security modules to manage encryption keys.
• The infrastructure makes use of AWS GuardDuty, Inspector, and Managed Container Services.
• Third-party security experts conduct regular penetration tests on the code, running applications, and deployment environments.
• The development process makes use of security tools such as CodeQL, Secrets Scanner, and Dependabot.
• The sharing of personal sensitive information with external parties usually requires explicit consent, except in cases governed by law or for security reasons.
• Non-identifiable data may be shared with related parties such as payment providers, healthcare providers, hospitals, and marketing agencies.
• Users can request to update or delete their personal information; to delete an unregistered session, session identification details are required.
• Users under 16 are not the target audience for public websites, and they should not submit personal information on their own.
• Do not send health, payment, or other sensitive information via regular email.

Safety checklist for medical institution procurement

1. Confirm Clearstep’s role in the contract, the BAA, the location of the data, and the sub-processors.
2. List all fields of protected health information and the data flow for each system.
3. Verify the methods for static storage, data transmission, backup, and key management.
4. Specify the deadlines for current penetration testing, audit reports, and vulnerability remediation.
5. Configure SSO, minimum permissions, logging, retention periods, and the recovery of accounts for employees who have left the company.
6. Processes for exporting and deleting test data, terminating contracts, and disaster recovery.
7. Establish clinical approval and change records for model or protocol updates.

Basic information

Recommendation score

It receives a rating of 4.3 out of 5 points. Clearstep integrates clinical triage protocols, chat and voice functionality, patient navigation, scheduling, and capacity management – making it suitable for American medical systems that need to restructure their patient access processes.

The main limitations are the lack of public prices and development documentation; to determine the accuracy of the figures, the purchaser must conduct a thorough analysis. The privacy policy also indicates that detailed health information will be collected. Individual users should not use it as a substitute for a doctor or as a tool for independent diagnosis.

Frequently Asked Questions

Can Clearstep diagnose diseases?

No. It provides support for symptom triage and decision-making regarding the path to seek medical care; the official terms clearly state that it cannot replace professional medical advice, diagnosis, treatment, or a thorough evaluation by a doctor.

Is Clearstep suitable for direct use by individuals?

It is primarily integrated into the digital platforms of hospitals or health systems, and there is no public option for individual subscriptions. Patients should use the access points provided by the partnering medical institutions and follow local medical recommendations.

How does Clearstep carry out triage?

The patient describes their symptoms and answers about 10 to 15 follow-up questions; the system then uses the Schmitt clinical protocol to assess the urgency of the situation, after which it directs the patient to emergency care, urgent care services, primary care facilities, specialized clinics, telemedicine, or home care.

What is the accuracy rate of Clearstep?

The current official website states a clinical accuracy rate of over 97%, while earlier sources indicated a rate of over 95%. It is necessary for organizations to take into account the specific research population, the gold standard used, the characteristics of severe cases, as well as any external validations; aggregate figures should not be regarded as guarantees for individual cases.

Does Clearstep support voice?

AI Voice is supported and can be used in nurse hotlines and contact centers. In cases of failed voice recognition, noise, accents, or dangerous signals, it is necessary to have a clear procedure for manual intervention.

Is it compatible with Epic and Cerner?

The official website lists integrations such as Epic, Oracle Health Cerner, and MEDITECH, and it is possible to connect to data related to facilities, inventory, scheduling, EMR, and CRM. The actual interfaces, versions, and costs are determined based on the specific implementation project.

How much is Clearstep?

There are no standard public prices; medical institutions need to schedule a demonstration in order to obtain a customized quote. Factors such as modules, the number of patients, voice functionality, integration options, security measures, and support services all influence the contract terms.

Does Clearstep offer a public API?

The product supports API integration for enterprises, but no public development documentation, self-service key options, or unified pricing for calls are available. Developers must obtain the interface specifications from the procurement and implementation team.

Is Clearstep open source?

It is not open source; no official source code repository or open-source license has been found. The details of the clinical agreement channels are publicly available, but this does not mean that Clearstep’s source code is made available to the public.

How will health data be used?

The platform collects symptoms, demographic information, and medical history in order to provide recommendations; it can also use aggregated, anonymized data to analyze behaviors and improve matching algorithms. The deployment of such platforms in specific medical institutions is subject to contracts, BAA agreements, and institutional privacy policies.

What should be done in case of emergency symptoms?

Contact the local emergency services or a doctor immediately; do not wait for Clearstep to complete the questions. Any AI symptom checker may misinterpret the input or fail to detect emergencies.

Summary

The core value of Clearstep is to transform patients’ symptoms into a safe, priority-based route to medical care, while also connecting them with doctors, locations, appointment services, and contact centers within hospitals. It is an infrastructure for medical institutions to use, rather than a tool for consumers to conduct diagnoses.

When selecting a medical facility, clinical validation, manual upgrades, deviation monitoring, system integration, and data management must take precedence over functional demonstrations. For patients using such facilities, doctors’ judgment and emergency assistance channels should always be available.