Higher Ed HX Solution | New Process Doubles Transcript Processing Efficiency
Hyland's Brainware Transcripts integrates with the Student Information System, but an outdated UI and low performance are causing decline in customer retention. We aim to use the Hyland Experience Platform(HxP) to enhance performance, boost revenue, and stand out in the market.
Project Tags
Deliverables
Design Audit
Visual Design
Competitor Analysis
User Research
User Flows
Design Presentation
User Testing
Year
2024
Team
1 UX Designer
1 Design Manager
1 User Researcher
2 Solution Consultants
2 Product Managers
Duration
4 months
Role
Product Designer
IMPACT 💸
Customers found the new experience to be more efficient and easier to use
Testing and Results
We presented the prototypes to six industry specialists who focus on higher education tasks.
All of them easily understood the new process of transcript processing.
They liked the keyboard shortcuts since it made their tasks more efficient.
Fig. A customer testing the prototype and providing feedback.
Personal Takeaways
I had the opportunity to speak with industry specialists (also called Evaluators) working in universities and high schools, and I discovered a whole new set of challenges they face.
I learned that complex processes can be simplified. The old process involved numerous steps, many of which I was able to consolidate into a single step.
All testing participants were excited about the launch of this solution. Brainware customers found the HX solution more usable and efficient for their daily tasks.
PROBLEM
Evaluators are frustrated with slow rate of Transcript Processing
Outdated UI
“The system process is complex, and the outdated interface makes it harder to learn the system quickly”
Performance Issues
“Data Verification is slow due to the model’s low performance and the need to verify numerous text fields manually”
Fig. The above images show the outdated and confusing UI of the current solution in Brainware.
Why do we need to fix this?
Business Goals
Increase customer retention
User Goals
Enhance transcript processing speed to minimize user corrections of incorrectly extracted data by improving model performance.
Increase revenue streams from building and selling a new solution in HX
User Interviews Helped Us Identify Inefficiencies in the Process Hindering Evaluators' Performance
A researcher was brought onto the project to conduct 3 user interviews with customers using the Brainware application, as well as 1 interview with an external customer. The goal was to understand their workflows, identify what was working well, and pinpoint areas for improvement.
Fig. A user researcher from my team interviewing a Brainware customer.
Key Insights
The outdated UI creates challenges in onboarding new evaluators efficiently.
Low model performance significantly increases the time required to verify extracted data, with some transcripts taking over 20 minutes to complete.
Repetitive steps in the process add unnecessary complexity and effort.
A lack of helpful analytics and insights makes it difficult for evaluators to track their progress and performance effectively.
Current Process Insight - Inefficient Workflow Expands Simple Tasks into Multiple Steps
Many repetitive steps can be condensed into fewer, more efficient ones. By rearranging the workflow, we can enhance overall efficiency. For example, after data extraction, evaluators can select the student, institution, and grades, then verify the document in an organized state, eliminating the need for an additional E-transcript step.
Fig. Illustrating old Workflow for Transcript processing that is hard for evaluators to follow. Highlighting that step 5 and 6 can be done in a single step to increase efficiency
SOLUTION
Simplified Workflow that Doubles Efficiency
Fig. Optimizing Workflow for Transcript processing that is easy for evaluators to follow
Enhancing Evaluator Experience through Personalized Insights and Gamification
After consulting with industry specialists, I incorporated metrics they identified as most valuable.
Today’s Goal Widget: This tool helps evaluators set achievable goals for the day. Given the potentially overwhelming number of transcripts, this feature creates clear expectations and provides the team with a sense of accomplishment.
Awards Widget: This feature is designed to boost morale, as the work can be repetitive and mundane. Recognizing achievements can help keep motivation high.
Processing Time Breakdown: This breakdown will allow evaluators to discuss their efficiency with managers, especially since their productivity is often influenced by the quality of the transcripts they receive
Matching student with Data Extracted from the Transcript
Since students with the same name often exist, we needed this screen. Typically, name, SSN, and date of birth are used to identify students, but industry specialists suggested that including an address could also be helpful for matching students in the student information system.
Streamlining Manual Data Verification Process
After the system has completed OCR and data extraction, some fields may not be accurately read. Therefore, evaluators need to manually verify these fields.
Color-coded keys match the type of information and data extracted: purple for student information and yellow for student records.
Fields that are matched form the student information system cannot be edited.
“Save for later” and “Submit” options provided. Submit button will be active when all the low confidence items are resolved.
4. The system displays the number of low-confidence items to inform the user which data items need manual verification. High-confidence fields do not require verification, thus speeding up the data verification process.
5. The user can also press the "tab" key to navigate to the next low-confidence data item, which is highlighted to provide visual feedback.
6. The semester table data can be collapsed. The "Delete Row" option becomes visible when one or more rows are selected.
Manager Dashboard: Filter Evaluator Performance, Track Progress Over Time, and Compare Model Performance Across Different Time Periods
Managers can view each evaluator's performance using time filters and discuss their progress with them. They can also monitor the types of incoming transcripts. Additionally, managers can track accuracy trends, allowing them to engage with machine learning engineers if the model's accuracy declines over time.
Next Steps
Streamline the process for transcript equivalency in the student portal.
Integrate the solution into the university registrar portal, enabling transfer students to easily check which credits will be transferred based on the transcript equivalency concept.
Next Project
