Blees.AI

Anchor 1

Blees.ai is a platform created for mass essay grading using Artificial Intelligence (AI) models.

The Blees.AI was a team effort with four members involved. While all of us worked on all aspects of the project, we each had our own focus.

My focus was on Research and Prototyping.

Brief

Blees.AI's goal:

To reduce the time and cost of mass grading essays.

Target users:

Certification bodies (CPA, CIMA etc.)

Educational Institutions (specifically post-secondary)

Deliverables:

4 high-fidelity wireframes for each section of the AI model.

Research

Testing

Prototypes

Research methods used and their key findings:

User Interviews:

Users needed some extra guidance while performing some tasks.
Users highlighted the importance of ‘custom feedback’ when grading, and how this feature should be integrated into the platform as it in integrated to their regular grading process.

Competitive Analysis

Upon inspecting Google Cloud ML, we were able to pinpoint specific features, such as the "Import" function, that could be improved and/or replicated in Blees.AI.

Heuristic analysis:

Strengths: minimal design, user control and consistency.
Improvements: error prevention, visibility of the system, flexibility and efficiency, and help and documentation .

The test was conducted using Maze and based on a navigation approach. This meant asking users to perform a specific task and measure their success. In addition to this, users were provided an opportunity at the end to provide feedback.

The strongest task success was on the Finding The Data Set and
Viewing Training Items pages. While the weakest task success was on the remaining pages.

Feedback provided: more guidance and direction needed about what the next steps are in the process.

There were many types of prototypes created. Sketches, low, mid and high fidelity wireframes.

The final prototype created using Figma included an interactive prototype. To go to the final design click here.