top of page
VA Thumbnail.png

Smart Health VA

Cut 40% Off Patient Intake Time with AI Virtual Assistant (VA)

INTRODUCTION

Role

UX Consultant

Timeline

Jan 2024 - May 2024

Tools

Figma, Miro, Flowise, LangSmith

Client

Highland Rivers Behavior Health (HRBH), one of Georgia's largest behavioral healthcare providers

Our Team

Saurabh, CEO

Rakesh, UX Director

Problem

How might we create an AI-driven solution to
streamline patient pre-consultancy and scheduling with emotional support?

Results

This project is currently in progress and has moved into the implementation phase. We are collaborating closely with HRBH staff and the IT team to ensure seamless integration and effective deployment of the AI-driven solutions.

Business Impact: We anticipate a significant reduction in administrative workload, greatly increasing revenue and improving operational efficiency.
Enhanced Patient Experience: The Virtual Assistant(VA) is designed to streamline the scheduling process and provide emotional support, potentially leading to huge reduction in patient onboarding time, higher patient satisfaction and reduced frustration during pre-consultation. 

  • 40% reduction in patient onboarding time 

  • Save up to 260,000 billable hours for staff

  • $15 million saved per year across the organization

Design

Approach

1 Research

Stakeholder Workshops

Diagram & Affinity Mapping


Pain Point


Persona

Problem Statement

2 Design

Solutions

Conversation Flow Chart
Conversation VA Design
Mid-fi Prototypes

3 Prototyping

Early Prototypes

Stakeholder Feedback

Prompt Testing & Iteration

My Contribution

  • Set design strategy with CEO and UX Director to align project vision with business goals

  • Created tailored AI solutions for new patients at HRBH

  • Designed conversational flows and interfaces from zero

  • Facilitated client workshops to dissect workflow inefficiencies and identify pain points

  • Managed client communications, delivering regular status and feasibility updates

  • Developed marketing materials to expand our company's market position

My role:

  • UX Consultant at The Generative AI Company LLC (TGAIC),  a design & consultancy startup, specializing in integrating SaaS AI solutions to organizations. I worked with Saurabh, the CEO and Rakesh, the UX Director.

 01 RESEARCH

Context

Highland Rivers Behavioral Health (HRBH) is one of Georgia's largest behavioral healthcare providers, serving over 25,000 clients annually across a 13-county region.

As the UX Consultant for this project, I played a pivotal role in shaping the design strategy and influencing key decision-making processes. By facilitating client workshops, conducting in-depth user research, and developing AI-driven solutions, I ensured our approach was aligned with HRBH's goals and needs, ultimately streamlining their patient pre-consultancy and scheduling process.

Problem Finding

At Highland Rivers Behavioral Health (HRBH), a significant challenge was identified: clinicians are spending 15 out of 40 working hours per week on administrative tasks, such as patient pre-consultancy and scheduling. This inefficiency resulted in thousands of work hours wasted and significant financial loss, along with low patient satisfaction. With hundreds of clinicians affected, the need for a streamlined solution became evident.

on non-billable administrative tasks for staff

40% of work time

patient satifactory

Extremely low

Research Constraints & Approaches

Conducting user interviews with frontline staff and patients at Highland Rivers Behavioral Health (HRBH) posed several challenges. The staff's busy schedules made it difficult to coordinate interviews without disrupting their essential work. Additionally, management preferred to limit frontline staff's awareness of potential changes at this early stage.

To overcome these constraints, I adopted the following two approaches: in-depth research and conducting workshops with the management team to gain deep insights into their operational challenges.

Approach 1:
In-Depth Research

  • Website and Annual Reports: I conducted thorough research using HRBH’s website, annual reports, and other publicly available documents to gather comprehensive insights into their operations and challenges.

  • Industry Reports: I utilized industry reports and studies on patient scheduling and consultation challenges in healthcare to gather relevant insights.

Approach 2:
Workshops With The Management Team

  • I organized 2 workshops with management team members to identify their most pressing concerns. These sessions enabled us to dive deep into each problem and dissect the current workflow, providing a clear picture of their operational landscape without involving frontline staff and patients at this stage.

  • I also gathered indirect frontline feedback by asking the management team specific questions about their observations and challenges of frontline staff.

Below is a screenshot of the workshop plan that I prepared and organized, targeting management members to efficiently identify and address HRBH's key challenges within a limited timeframe.

Pain Points in Patient Pre-consultancy & Scheduling

During the workshops, I documented the challenges shared by the management team at each step of the current workflow. Based on that, I organized these into key aspects and presented them to the CEO and UX Director for discussion. Here are the major issues we agreed on to prioritize.

1. Repeated Work for Staff

Staff members face redundancy with initial consultations and data input into the EMR system.

2. Patient Frustration on Filling Complicated Form 

Patients often become impatient and frustrated, leading to errors during the form-filling process.

3. Unidentified Emotional Needs

Patients frequently do not recognize or disclose their mental health needs upfront.

4. Lack of Information on Additional Resources

Patients lack access to information about other vital resources, such as food centers.

Persona

Based on thorough research, I developed 2 personas and sketched out their story, addressing the key pain points, using a new tool Storytribe. Below is the story of the primary persona, a potential patient, Sofia Martinez.

Sofia Martinez

Female, 37 years old
Family: An immigrant single mother of two children(ages 8 and 10)
Education: High school graduate, attended some college courses
Employment: Works as a cashier at a local grocery store

01 Overwhelmed by life pressure

As an immigrant single mother, she has a stressful life, her chronic migraines and depression make it worse.

03 Struggling with Language Barrier

Sofia calls HRBH hoping to know more but struggles with the language barrier.

02 Too much online information

Searching for help, Sofia finds HRBH but the abundance of information makes it hard to read through.

04 Frustrated by Long Forms 

Sofia finds the in-take form sent from HRBH long and confusing, making her more frustrated. Finally, she gives up.

Problem Statement

How might we create an AI-driven solution to
streamline patient pre-consultancy and scheduling with emotional support?

02 Solution & Design

Our Solution: Intelligent Virtual Assistant (VA) for Patient Pre-Consultancy & Scheduling

Based on discussions with the CEO and UX Director, I developed an Intelligent Virtual Assistant (VA) aimed at streamlining the pre-consultancy and scheduling processes. This VA significantly reduces the administrative burden on clinicians and enhances patient satisfaction by providing a user-friendly and efficient solution. Below are the five key features of our VA that make this possible.

Intuitive Interface

Ensure an easy-to-navigate interface tailored to Sofia's tech comfort levels and needs

Natural and Multilingual

Use simple, easy-to-understand language with multilingual options to make interactions smoother for Sofia

Emotional Reassurance

Offer emotional support and reassurance during interactions to help Sofia feel supported

Comprehensive Resources

Provide information on community resources like food centers and support groups

Multi-Channel Access

Enable 24/7 communication via phone, email, and the HRBH website for accessibility

Virtual Assistant(VA) Characteristics

Based on the proposed solution, I identified and developed several key characteristics for the Intelligent Virtual Assistant. After thorough discussions with the UX Director, we refined these characteristics to ensure they meet both patient and staff needs. Here are the three primary characteristics of our VA.

Approachable & Clear
Supportive & Empathetic
Efficient & Trustworthy

Conversation Flow Chart

Based on the solution and with the VA characteristics and HRBH’s current processes in mind, I created a detailed conversation flow chart below. This flowchart outlines the interaction between the VA and potential patients during pre-consultancy and scheduling. It includes clear visual indicators for each step from start to end. This structured approach ensures a smooth and user-friendly experience, laying the groundwork for future prototyping and development.

03 Prototyping

VA Early Prototyping

To demonstrate AI's capabilities to the HRBH management team, I helped create a low-fidelity VA using Flowise. I developed an example conversation script based on the flowchart, incorporating emotional support elements to enhance patient interactions. This script was crucial in guiding the prototype development, ensuring the VA could address patient needs effectively while providing compassionate support. Below is one example of pre-consultancy with VA.

Success Metrics

To measure the effectiveness of the Intelligent VA from users' standpoint, I identified three key success metrics:

  1. Reduction in Patient Intake Time: Aiming for a 40% decrease in hours spent on pre-consultancy and scheduling activities.

  2. Improved Patient Satisfaction: Aiming to achieve a 30% increase in patient satisfaction scores, as measured by post-visit surveys and feedback forms.

  3. Increase in Appointment Completion Rates: Target a 25% increase in the appointment completion rate, which indicates the VA’s efficiency in scheduling and maintaining patient engagement throughout the process.

The metrics will help us evaluate the VA's impact and ensure to meet the needs of staff and patients effectively.

Additionally, I also created additional metrics focused on evaluating the VA’s performance, as shown in the table.

Testing & Prompt Iteration 

Given our limited time and resources, we couldn't conduct extensive user testing. However, I invited a few volunteers to test the VA. To demonstrate our ability to monitor and modify the VA post-deployment, here are my approaches. 

Self-Learning LangSmith

I took the initiative to self-learn LangSmith, a tool primarily used for engineering tasks, to monitor and modify the VA through simple prompt changes.

Mind Map Creation

I created a detailed mind map to outline the key features of  LangSmith to monitor and modify the VA effectively.

Observing User Behavior

During the testing phase, I noticed instances of hesitancy and a desire to abandon the conversation at certain points. So I asked following-up questions to improve the VA.

Prompt Modification

Patients lack access to information about other vital resources, such as food centers.

LangSmith User Guide Presentation

I developed a LangSmith user guide presentation showcasing this specific case to demonstrate to our CEO and the client how we can continuously improve the VA based on user feedback.

04 Results & Impact

Stakeholder Feedback

The HRBH management team provided overwhelmingly positive feedback on our prototypes, praising both the innovative designs and our clear, consistent communication throughout the project. They highlighted how our proactive attitude and collaborative approach significantly contributed to the project's success. My efforts in facilitating workshops, conducting thorough research, and developing user-centric AI solutions were instrumental in achieving this positive reception.

These enhancements are expected to improve overall patient satisfaction, setting a new standard for how AI can optimize pre-consultancy and schduling in the healthcare industry.

The estimated impact...

for patient pre-consultancy and scheduling 

40% time reduction

for staff on patient intake

260,000 hours saved

in annual saving

$15 million

Next Step

User Research with Patients & Frontline Staff

Conduct more in-depth user research sessions with patients and frontline staff to gather insights on their specific needs and challenges. This will ensure the VA is finely tuned to their day-to-day operations and challenges.

VA Conversation Design and Hi-Fi Prototype

Refine the VA conversation design to make interactions even more seamless and intuitive. Develop a high-fidelity prototype that incorporates all feedback and improvements, providing a near-final version of the VA for final testing and validation.

Development Integration

Develop and integrate the AI solution within HRBH’s current official website and EMR system. This step involves close collaboration with their IT department to ensure seamless integration and minimal disruption to their existing workflows.

bottom of page