AI
Optical Character Recognition
Fax processing
In U.S. healthcare, medical service requests typically arrive via faxes. An intake user processes these by verifying member eligibility, benefits and existing requests to check for duplicates. If a new service is needed, the intake user initiates a ‘create service’ shell. The service request notification is then communicated to the provider or patient via phone, fax, mail, or portal.
Role
Senior Product (UX) Designer
Client
HealthEdge
Year
2023
Timeline
6 Months
Final Product
Problem
This process is highly manual - intake users must read lengthy forms and manually enter data, making it time-consuming and error-prone. These inefficiencies delay patient service allocation, disrupt invoicing, and extend turnaround time, impacting both care delivery and business operations.
Our Goal
Use AI/OCR technology to automate fax processing, extract data, provide smart suggestions, detect duplicates, and streamline workflows.
What is AI-powered Optical Character Recognition?
OCR (Optical Character Recognition) extracts data from scanned documents, PDFs, and images, eliminating manual entry. AI-powered OCR improves accuracy with computer vision and machine learning, while also suggesting actions for users.
How does it help users?
✅ Data Extraction: Saves time and effort.
⚙️ Automated Processes: Reduces errors and enhances accuracy.
📈 Increased Productivity: Enables faster service shell creation.
Design Process
A research-driven approach to analyze the current intake process and evaluate AI capabilities.
Internal Workshops
The Leadership, Product Owner, and Design Director discussed client-facing challenges in the intake workflow, helping us identify key pain points. These firsthand insights guided us in preparing a workshop for customers (users). We also explored AI capabilities to address these challenges and enhance user experience.
Additionally, we considered hiring a vendor partner to support AI/OCR implementation, which helped us prepare for vendor workshops. As a first step, I did extensive research on OCR technology and AI, which was instrumental in shaping my approach to solving this problem.
Customer Workshops
During the customer workshops, we brainstormed with users and asked them to write down what would help them complete long forms more easily. This exercise gave us valuable insights into user expectations.
We then discussed their current pain points while working on intake forms, allowing us to better understand their challenges and refine our approach accordingly.
Vendor Workshops
We had multiple vendor workshops where they showcased their real-time products and provided insights into AI accuracy and data extraction. These sessions helped us understand how their solutions could be integrated as a third-party tool for OCR extraction.
However, the idea was ultimately rejected, as our goal was to build a long-term, scalable solution that could be white-labeled for other clients. As these sessions progressed, we categorized the AI functions and intake categories, that was needed to build the solution.
02 / Research Findings
Based on all the workshops, we defined UX goals and provided recommendations, ensuring alignment with user needs by validating them directly with users. Once our goals were set, we quickly updated the persona and user flows. We focused heavily on low-fidelity designs, gathering constant feedback to iterate rapidly and refine the solution.
UX Goals
Simplify Workflows
Reduce complexity and streamline workflows to minimize user effort and reduce errors.
Enhance Automation and Integration
Leverage automation and better integrate tools and data sources to support more efficient and accurate decision-making.
Improve User Experience
Focus on improving the user interface and experience, providing better tools, real-time information, and support to enhance overall satisfaction and productivity.
Increase Flexibility and Support for Complex Cases
Ensure the system can handle a variety of case complexities with ease, reducing the burden on users and improving outcomes.
UX Recommendations
Enhance Document Handling and Viewing
Improve the document viewer to ensure consistent functionality. Enable a reliable multi-document interface that supports side-by-side viewing and easy navigation between documents to facilitate better review processes.
Streamline Data Entry and Management
Integrate Optical Character Recognition (OCR) and Natural Language Processing (NLP) tools to automatically extract and populate data from fax and email submissions directly into the system.
Enhance Decision Support Tools
Develop APIs or direct integrations with external decision support tools like MCG to automatically populate data and streamline the decision-making process. This integration should allow users to access these tools within the GuidingCare interface without manual input.
Persona
Gap Analysis
Cluttered Dashboard
Users struggle to identify and prioritize faxes due to unclear categorization.
There's no clear indication of urgency, making it difficult to determine which fax to address first.
Faxes are displayed in a small widget with no visual cues for new or high-priority items.
Users go through multiple steps - selecting the module, waiting for data to load, selecting a checkbox, and clicking the view icon - just to access a fax.
The "View Fax" icon does not clearly indicate that it’s the action needed to work on a fax.
Outdated and interruptive user experience
The fax opens above the workspace, creating a poor user experience. Users cannot view both widgets simultaneously, forcing them to constantly minimize and expand panes, which disrupts workflow and increases cognitive load.
Too much manual effort
The user has to manually search for the member, read the fax, and analyze whether the identified member is correct. Additionally, they must check existing authorizations separately, navigate to a different page to review details, and then determine if it is a duplicate, making the process time-consuming.
Information Flow
Layout Explorations
Final Designs
Extensive collaboration with cross-functional teams, vendor and leadership
A strong grip on refining and executing product strategy
Working with the users and data to continuously iterate the design
Understanding the technicalities of working wIth Al & OCR