Hey...
assess
assess
assess
assess
AI-Powered Education Platform for Arvion State University
AI-Powered Education Platform for Arvion State University
AI-Powered Education Platform for Arvion State University
AI-Powered Education Platform for Arvion State University
asses
asses
the future of grading starts here
the future of grading starts here
In 2022, CSU partnered with our team to build an AI-powered education platform that didn’t just automate grunt work—it redefined teaching and learning which help them to reducing the grading time by 60% and improve the user efficiency by 78%. From battling skepticism ("Will robots replace professors?") to triumphant pilots ("I got my grade before my next class!"), here’s how we transformed frustration into innovation.
In 2022, CSU partnered with our team to build an AI-powered education platform that didn’t just automate grunt work—it redefined teaching and learning which help them to reducing the grading time by 60% and improve the user efficiency by 78%. From battling skepticism ("Will robots replace professors?") to triumphant pilots ("I got my grade before my next class!"), here’s how we transformed frustration into innovation.
In 2022, CSU partnered with our team to build an AI-powered education platform that didn’t just automate grunt work—it redefined teaching and learning which help them to reducing the grading time by 60% and improve the user efficiency by 78%. From battling skepticism ("Will robots replace professors?") to triumphant pilots ("I got my grade before my next class!"), here’s how we transformed frustration into innovation.
In 2022, CSU partnered with our team to build an AI-powered education platform that didn’t just automate grunt work—it redefined teaching and learning which help them to reducing the grading time by 60% and improve the user efficiency by 78%. From battling skepticism ("Will robots replace professors?") to triumphant pilots ("I got my grade before my next class!"), here’s how we transformed frustration into innovation.
Client:
assess
My Role:
Product Designer
Year:
2022
Service Provided:
Product Design, Web Design, Mobile Design
Vision
Vision
Enhance Scalability by supporting large classes without compromising feedback quality.
Reduce dependency on TAs for grading.
Track student performance trends for accreditation reporting and Identify class-wide knowledge gaps.
Lower reliance on third-party tools (e.g., Gradescope, Turnitin).
Hybrid Learning Needs to support futuristic education technologies or next gen courses
Vision
Enhance Scalability by supporting large classes without compromising feedback quality.
Reduce dependency on TAs for grading.
Track student performance trends for accreditation reporting and Identify class-wide knowledge gaps.
Lower reliance on third-party tools (e.g., Gradescope, Turnitin).
Hybrid Learning Needs to support futuristic education technologies or next gen courses
My Responsibility
Leading the Product Design team.
Define Product Strategy with Marketing lead: Align the AI service with business goals
User Research Oversight (Planning & execution): Validate core problems with teachers/students
Stakeholder Collaboration: Bridge gaps between VX team, developers, QA, client, and legal.
My Responsibility
Leading the Product Design team.
Define Product Strategy with Marketing lead: Align the AI service with business goals
User Research Oversight (Planning & execution): Validate core problems with teachers/students
Stakeholder Collaboration: Bridge gaps between VX team, developers, QA, client, and legal.
My Responsibility
Leading the Product Design team.
Define Product Strategy with Marketing lead: Align the AI service with business goals
User Research Oversight (Planning & execution): Validate core problems with teachers/students
Stakeholder Collaboration: Bridge gaps between VX team, developers, QA, client, and legal.
Challenges
Challenges
Challenges
Technical Limits - Handwritten answers, textbook image errors, or complex subjects (like essays) confusing the AI.
Scalability - Discipline-specific needs
User Trust - Stakeholders doubt on AI’s accuracy and fear of unfair grading.
Workflow Fit - Slotting the tool smoothly into existing workflow of teachers.
Privacy & Rules - Protecting student data (like answer sheets) and following university policies (FERPA).
Technical Limits - Handwritten answers, textbook image errors, or complex subjects (like essays) confusing the AI.
Scalability - Discipline-specific needs
User Trust - Stakeholders doubt on AI’s accuracy and fear of unfair grading.
Workflow Fit - Slotting the tool smoothly into existing workflow of teachers.
Privacy & Rules - Protecting student data (like answer sheets) and following university policies (FERPA).
Success Criteria
Success Criteria
Accuracy & Performance: AI grading must match professor scores ≥90% of the time and generate usable quiz questions 85% of the time.
Efficiency Gains: Cut grading time by ≥60% and quiz creation to ≤30 minutes per exam.
User Adoption & Trust: Achieve ≥75% professor adoption and ≥80% student satisfaction with feedback speed.
Accuracy & Performance: AI grading must match professor scores ≥90% of the time and generate usable quiz questions 85% of the time.
Efficiency Gains: Cut grading time by ≥60% and quiz creation to ≤30 minutes per exam.
User Adoption & Trust: Achieve ≥75%
Compliance & Ethics: Maintain 100% FERPA/GDPR compliance and <3% grading bias across demographics.
Technical Reliability: Ensure 99.9% system up-time and 95% mobile upload success.
Institutional Impact: Reduce paid tool dependency by 50% and support hybrid/online courses equally.
professor adoption and ≥80% student satisfaction with feedback speed.
Compliance & Ethics: Maintain 100% FERPA/GDPR compliance and <3% grading bias across demographics.
Technical Reliability: Ensure 99.9% system up-time and 95% mobile upload success.
Institutional Impact: Reduce paid tool dependency by 50% and support hybrid/online courses equally.
Accuracy & Performance: AI grading must match professor scores ≥90% of the time and generate usable quiz questions 85% of the time.
Efficiency Gains: Cut grading time by ≥60% and quiz creation to ≤30 minutes per exam.
User Adoption & Trust: Achieve ≥75% professor adoption and ≥80% student satisfaction with feedback speed.
Compliance & Ethics: Maintain 100% FERPA/GDPR compliance and <3% grading bias across
demographics.
Technical Reliability: Ensure 99.9% system up-time and 95% mobile upload success.
Institutional Impact: Reduce paid tool dependency by 50% and support hybrid/online courses equally.
Success Criteria
UX Design Strategy
UX Design Strategy
UX Design Strategy
Since we had to be working with a fixed deadline for the Alpha/Phase1 release, we use the Double Diamond + MoSCoW approach.
Starting with Double Diamond :
Since we had to be working with a fixed deadline for the Alpha/Phase1 release, we use the Double Diamond + MoSCoW approach.
Starting with Double Diamond :
PROBLEM SPACE
DISCOVER
DIVERGENCE
Mind Map
Brainstorming
User Interview
User Persona
Competitive Analysis
Journey Mapping
Problem Definition
User Flow
Wireframing
Moodboard
Wireframing
Design screens
Prototyping
MoSCoW Plan
Phased Approach
Core Features
Usability Testing
DIVERGENCE
DEVELOP
DEFINE
CONVERGENCE
CONVERGENCE
DELIVER
SOLUTION SPACE
Discover Phase (Research & Problem Definition)
Given that we have precisely identified our application's target audience, we will conduct in-depth user research in the Discover phase, leveraging methods such as stakeholder interviews, competitive analysis and persona development to uncover key needs and pain points.
Given that we have precisely identified our application's target audience, we will conduct in-depth user research in the Discover phase, leveraging methods such as stakeholder interviews, competitive analysis and persona development to uncover key needs and pain points.
User Interview
User Interview
User Interview
Given that we have precisely identified our application's target audience, we will conduct in-depth user research in the Discover phase, leveraging methods such as stakeholder interviews, competitive analysis and persona development to uncover key needs and pain points.
Discover Phase (Research & Problem Definition)
Discover Phase (Research & Problem Definition)
Based on the requirements outlined by the Arizona University team, we developed a user interview questionnaire to gather insights. For this study, we selected 20 participants from ASU, consisting of 9 professors, 9 students, and 2 additional participants for a mock interview to refine the questions. All
Based on the requirements outlined by the Arizona University team, we developed a user interview questionnaire to gather insights. For this study, we selected 20 participants from ASU, consisting of 9 professors, 9 students, and 2 additional participants for a mock interview to refine the questions. All participants have expressed their willingness to share their perspectives, pain points, and motivations.
participants have expressed their willingness to share their perspectives, pain points, and motivations.
Based on the requirements outlined by the Arizona University team, we developed a user interview questionnaire to gather insights. For this study, we selected 20 participants from ASU, consisting of 9 professors, 9 students, and 2 additional participants for a mock interview to refine the questions. All participants have expressed their willingness to share their perspectives, pain points, and motivations.
Criteria considered for selecting interview participants:
Diversity in Degree Programs – Covered varied use cases (STEM, Humanities, etc).
Representation of Class Sizes – Small & large lectures having different needs.
Balanced Tech Literacy – Avoided bias by including both tech-savvy and hesitant users.
Criteria considered for selecting interview participants:
Diversity in Degree Programs – Covered varied use cases (STEM, Humanities, etc).
Representation of Class Sizes – Small & large lectures having different needs.
Balanced Tech Literacy – Avoided bias by including both tech-savvy and hesitant users.
User Interview Approach : Professors
User Interview Approach : Professors
Goal: To understand their workflow, pain points, and expectations from an AI grading & course management tool. The Questionnaire is categorized as below :
Goal: To understand their workflow, pain points, and expectations from an AI grading & course management tool. The Questionnaire is categorized as below :
Current Teaching & Grading Process
AI-Assisted Grading & Assignment Generation
Course & Assignment Creation
Concerns & Adoption
Current Teaching & Grading Process
AI-Assisted Grading & Assignment Generation
Course & Assignment Creation
Concerns & Adoption
User Interview Approach : Students
User Interview Approach : Students
Goal: To understand their study habits, feedback preferences, and concerns about AI grading. The Questionnaire is categorized as below :
Goal: To understand their study habits, feedback preferences, and concerns about AI grading. The Questionnaire is categorized as below :
Current Learning & Submission Process
Instant AI Grading & Feedback
AI-Generated Assignments & Study Help
Concerns & Trust in AI
Current Learning & Submission Process
Instant AI Grading & Feedback
AI-Generated Assignments & Study Help
Concerns & Trust in AI
Affinity Map
Affinity Map
Affinity Map
Result :
Professors are more cautious than students about adoption
Students care more about instant feedback while professors prioritize accuracy
STEM faculty are more open to AI grading than humanities faculty
Privacy concerns are significant for both groups
Grade appeal mechanisms are important for acceptance
Result :
Professors are more cautious than students about adoption
Students care more about instant feedback while professors prioritize accuracy
STEM faculty are more open to AI grading than humanities faculty
Privacy concerns are significant for both groups
Grade appeal mechanisms are important for acceptance
Result :
Professors are more cautious than students about adoption
Students care more about instant feedback while professors prioritize accuracy
STEM faculty are more open to AI grading than humanities faculty
Privacy concerns are significant for both groups
Grade appeal mechanisms are important for acceptance
Competitive Analysis
Competitive Analysis
Competitive Analysis
After prioritizing key features from user research, we benchmarked against existing AI-driven solutions in the market to assess competitive capabilities.
After prioritizing key features from user research, we benchmarked against existing AI-driven solutions in the market to assess competitive capabilities.
After prioritizing key features from user research, we benchmarked against existing AI-driven solutions in the market to assess competitive capabilities.
Market Gaps & Opportunities :
No single tool combines grading, assignment generation and course creation.
Handwritten answer sheet grading is still limited in most tools.
Few AI solutions offer seamless textbook-to-course conversion.
Market Gaps & Opportunities :
No single tool combines grading, assignment generation and course creation.
Handwritten answer sheet grading is still limited in most tools.
Few AI solutions offer seamless textbook-to-course conversion.
Market Gaps & Opportunities :
No single tool combines grading, assignment generation and course creation.
Handwritten answer sheet grading is still limited in most tools.
Few AI solutions offer seamless textbook-to-course conversion.
User Journey Map
User Journey Map
User Journey Map
As per the competitive analysis, a user experience mapping is created to understand the opportunities we have considering the existing similar functionality tools or services.
As per the competitive analysis, a user experience mapping is created to understand the opportunities we have considering the existing similar functionality tools or services.
Define Phase
In the Define phase of the Double Diamond framework, we synthesize research insights to clearly articulate the core problem—ensuring our solutions are grounded in real user needs before moving into development.
In the Define phase of the Double Diamond framework, we synthesize research insights to clearly articulate the core problem—ensuring our solutions are grounded in real user needs before moving into development.
In the Define phase of the Double Diamond framework, we synthesize research insights to clearly articulate the core problem—ensuring our solutions are grounded in real user needs before moving into development.
User Persona
User Persona
Based on insights from our primary user groups, we created two key personas: an Educator and a Student. These personas represent our core audience and will guide all design decisions for the platform.
Based on insights from our primary user groups, we created two key personas: an Educator and a Student. These personas represent our core audience and will guide all design decisions for the platform.
Based on insights from our primary user groups, we created two key personas: an Educator and a Student. These personas represent our core audience and will guide all design decisions for the platform.
PoV
PoV
Tutors need an LMS that automates course creation, lectures, assignments, and rubrics - including auto-grading with manual customization - to reduce backend workload while accommodating individual student needs and lecture adaptability.
Tutors need an LMS that automates course creation, lectures, assignments, and rubrics - including auto-grading with manual customization - to reduce backend workload while accommodating individual student needs and lecture adaptability.
Students require an LMS that automatically creates personalized mock tests and delivers instant, detailed feedback to help them study more efficiently.
Students require an LMS that automatically creates personalized mock tests and delivers instant, detailed feedback to help them study more efficiently.
Mission
Students
Faculty & Instructors
To transform education efficiency by reducing administrative burdens on instructors while enhancing learning outcomes for students—leveraging AI to automate time-consuming tasks like grading, assessment creation, and course structuring.
Save time & burnout by cut grading time.
Improve Consistency & Fairness by reducing human bias in grading
Streamline Course Development by converting textbook chapters/images into structured syllabi, lecture notes, and quizzes in minutes.
Instant, Actionable Feedback within minutes with explanations for mistakes
Reduce anxiety around grading delays and unclear feedback.
Mission
Students
Faculty & Instructors
To transform education efficiency by reducing administrative burdens on instructors while enhancing learning outcomes for students—leveraging AI to automate time-consuming tasks like grading, assessment creation, and course structuring.
Save time & burnout by cut grading time.
Improve Consistency & Fairness by reducing human bias in grading
Streamline Course Development by converting textbook chapters/images into structured syllabi, lecture notes, and quizzes in minutes.
Instant, Actionable Feedback within minutes with explanations for mistakes
Reduce anxiety around grading delays and unclear feedback.
Mission
Students
Faculty & Instructors
To transform education efficiency by reducing administrative burdens on instructors while enhancing learning outcomes for students—leveraging AI to automate time-consuming tasks like grading, assessment creation, and course structuring.
Save time & burnout by cut grading time.
Improve Consistency & Fairness by reducing human bias in grading
Streamline Course Development by converting textbook chapters/images into structured syllabi, lecture notes, and quizzes in minutes.
Instant, Actionable Feedback within minutes with explanations for mistakes
Reduce anxiety around grading delays and unclear feedback.
Develop Phase
The Develop phase blends structured planning with creative ideation—prototyping, testing, and iterating on solutions to ensure they align with both user needs and strategic goals.
The Develop phase blends structured planning with creative ideation—prototyping, testing, and iterating on solutions to ensure they align with both user needs and strategic goals.
Strategy
Strategy
Breaking the project into sequential phases (not sprints), each with a mini-deadline:
Breaking the project into sequential phases (not sprints), each with a mini-deadline:
Feature Prioritization (MoSCoW Method)
Categorized features based on CSU’s needs and deadline constraints:
Feature Prioritization (MoSCoW Method)
Categorized features based on CSU’s needs and deadline constraints:
User Flow
User Flow
Through user mapping analysis, we developed four core task. These informed the creation of a comprehensive user flow (both instructor & student), outlining all potential pathways and opportunities for user navigation.
Below you can find a basic user flow :
Through user mapping analysis, we developed four core task. These informed the creation of a comprehensive user flow (both instructor & student), outlining all potential pathways and opportunities for user navigation.
Below you can find a basic user flow :
Wireframing
Wireframing
100’s of layouts which need to be complementing with the existing LMS of CSU which also be providing a sophisticated and simpler user experience. We end up with few of the below provided layouts. Some mandatory step cannot be avoided, since it is much more flexible for creative process.
100’s of layouts which need to be complementing with the existing LMS of CSU which also be providing a sophisticated and simpler user experience. We end up with few of the below provided layouts. Some mandatory step cannot be avoided, since it is much more flexible for creative process.
Moodboard
Moodboard
Helvetica - the iconic modern typeface, to keep up the consistency with the existing brand products. This will also be the best for printing and digital functionality, which is an important use case we deal in the current product.
Helvetica - the iconic modern typeface, to keep up the consistency with the existing brand products. This will also be the best for printing and digital functionality, which is an important use case we deal in the current product.
Deliver Phase
The Develop phase blends structured planning with creative ideation—prototyping, testing, and iterating on solutions to ensure they align with both user needs and strategic goals.
The Develop phase blends structured planning with creative ideation—prototyping, testing, and iterating on solutions to ensure they align with both user needs and strategic goals.
User Interface
Usability Testing
We conducted usability testing with 8 participants selected based on their willingness and responses from preliminary user interviews. The testing included both remote moderated (live observation) and remote unmoderated (self-guided) sessions, focusing on qualitative feedback. To ensure comprehensive evaluation, we designed multiple real-world use case scenarios—such as instant grading, assignment generation, and course creation from textbook uploads—and recorded sessions for quantitative analysis (e.g., task completion rates, time-on-task, and error frequency).
Usability Testing
We conducted usability testing with 8 participants selected based on their willingness and responses from preliminary user interviews. The testing included both remote moderated (live observation) and remote unmoderated (self-guided) sessions, focusing on qualitative feedback. To ensure comprehensive evaluation, we designed multiple real-world use case scenarios—such as instant grading, assignment generation, and course creation from textbook uploads—and recorded sessions for quantitative analysis (e.g., task completion rates, time-on-task, and error frequency).
Usability Testing
We conducted usability testing with 8 participants selected based on their willingness and responses from preliminary user interviews. The testing included both remote moderated (live observation) and remote unmoderated (self-guided) sessions, focusing on qualitative feedback. To ensure comprehensive evaluation, we designed multiple real-world use case scenarios—such as instant grading, assignment generation, and course creation from textbook uploads—and recorded sessions for quantitative analysis (e.g., task completion rates, time-on-task, and error frequency).
Few of testers insights :
