Project
001
Minerva Studio @GEP
A self-serve platform that simplifies the development and usage of AI & ML use cases, accelerating data-driven decision-making.
This was my first project as an associate UX designer at GEP Worldwide. I worked on Minerva Studio, a self-serve AI & ML platform designed to simplify data-driven decision-making. My role involved designing key components and creating new features, making complex AI workflows accessible to non-technical users while preserving advanced functionality for data scientists. This project enhanced my skills in enterprise UX design and cross-functional collaboration. Due to an NDA, specific details and the full prototype cannot be displayed here.
Info
002
Role
UX Designer
Timeline
May 2022 — Sep 2023
Tools
Axure RP & Figma
Team
1 Senior Designer,
1 Product Manager,
5 Developers
Overview
003
Problem
GEP needed to make AI and ML capabilities more accessible across any organization and to the customers, but complex workflows and technical barriers were limiting adoption and slowing down data-driven decision-making processes.
Outcome
Minerva Studio provided organizations with a platform to harness AI and ML capabilities without requiring deep technical knowledge. By simplifying complex workflows and offering intuitive interfaces, we enabled customers to implement projects quickly and easily resulting in faster insights, improved decision-making processes,
Uncovering Issues
004
User research —
findings & insights
Before my onboarding, the team conducted a through research to gain insights into user needs and preferences, Here are the key findings from the research.
Users felt existing AI platforms lacked adequate guidance for beginners.
Many users felt lost when using AI tools, citing a steep learning curve and a lack of clear starting points for newcomers to AI and ML technologies.
Business users desired a 'no-code' approach to leveraging AI/ML
The research revealed a strong preference for simplified interfaces among non-technical users. Users expressed a desire to input data and receive actionable insights without needing to write code or understand complex algorithms.
Users found existing AI/ML platforms too complex for their skill level.
Many non-technical users struggled with implementing AI/ML solutions. The complexity of current tools was cited as a major barrier to adoption, with users feeling that extensive technical knowledge was required to utilize these platforms effectively.
Product Vision
005
Product Vision - key value proposition
The key value and vision of the product was to "Empower any organization to readily leverage GEP’s AI capabilities for their supply chains via a ‘No Code’ self serve platform and without the need of any specialized skills."
Initial Concept
006
Initial concept - ideas
Here are some snippets of the initial concept ideas of how the product should look and what all features it should have.
Design Evolution
009
Design Evolution - key considerations
Analysis of user feedback from the Minerva Studio MVP revealed key areas of opportunity, informing our iterative design process to further optimize the user experience. Thus based on the feedback following changes were made to the product.
Home Page
The user lands on this page immediately after logging in. The page provides an overview of all the current flows that are available, highlights any active flows, and offers some related resources to help users understand how to use the tool.
Models Page
After the landing page, users can either utilize a pre-existing trained model or train a new one. Machine learning models are broadly classified into Supervised and Unsupervised categories. To simplify these complex terms, we've created two distinct sections with clear descriptions of their functionality. This helps users easily differentiate and choose the most suitable model for their needs.
Orchestration Page
After training a model, users can design a flow to test new data on the orchestration page, which functions as a blank canvas for creating custom workflows. While an initial structure is provided, users have the flexibility to rearrange nodes and modify the flow as needed. Once created, this flow is conveniently accessible from the home page for direct use.
