As the volume of digital content grows, users increasingly struggle to efficiently organize and access their files. Traditional file management systems often rely on manual sorting, naming conventions, and folder structures, which can be time-consuming and prone to inconsistency. Our challenge is to design an AI-powered solution that helps users quickly and intuitively organize their files. The solution should enhance productivity, reduce cognitive load, and seamlessly integrate into existing workflows.
To address the challenge of file organization, we conducted a comprehensive competitive analysis to understand current solutions and identify any gaps. Leveraging these insights, we designed an AI-assisted file organization system that automatically categorizes and suggests file/folder names based on content. Through two iterative rounds of user testing, we refined the interface and interaction model to ensure the experience was intuitive, efficient, and aligned with real user needs. The result is a smart, user-centric tool that streamlines file organization and improves accessibility.
As an Associate Designer and Researcher on the AI team, I contributed to both the design process and user research. I was also responsible for maintaining clear and organized project documentation within FigJam, ensuring that meeting notes and research findings were easily accessible and up to date throughout the project lifecycle.
Over the course of 12 weeks, our team worked on designing the AI aspect of Trevevah. Once the initial user research and competitive analysis were completed, we transitioned into a structured design and test cycle. Every two weeks, we iterated through ideation, prototyping, and usability testing to refine our concepts based on user feedback.
As part of our primary research, we conducted a detailed competitive analysis to understand the landscape of both direct and indirect competitors. This exercise helped us identify best practices in the current market to better inform our design strategy.
Our direct competitors include Google Drive & Gemini, Microsoft OneDrive & CoPilot, Box AI, and DropBox Dash.
Our indirect competitors include Sparkle, Docupile, ClickUp AI, Tabbles, Zapier, and Visioneer AI Organizer.
We planned and created a survey to gain insights on how users current file and folder organization habits, challenges, and preferences, and to explore user interest in AI-powered file management. We aimed to gather both quantitative and qualitative data from a group that closely reflects our target user to identify opportunities and current challenges to guide the design of the Treevah’s AI-powered file organization.
The survey was shared through our personal network in person or through social network sites. Overall, we received 31 responses.
From this survey, we found that users want AI assistance but not full automation, preferring rule-based organization with previews and control over changes.
Based on the initial research and user survey, I designed a workflow centered around a a key use case: having AI assist the user in file rename and organization upon upload.
In this flow, the user starts by uploading files, which triggers a pop-up containing two columns to display the files. The left column shows the user how the files appear before renaming and organization while the right column shows how the files appear after the changes. Users select a button to have AI rename the files as many times as they want, with the option to manually rename files as needed.
Once naming is complete, the user proceeds to the organization step, where the AI automatically categorizes the files and suggests folder names. Users retain full control to adjust the organization manually, ensuring flexibility while still benefiting from AI-driven efficiency.
When the user selects "Complete", the pop-up closes and the newly uploaded files are visually highlighted across the three-panel layout to indicate their updated status.
As the project evolved, the design direction shifted to align with the product owner's vision of differentiating the platform through real-time, AI-assisted organization. This pivot required moving away from the initial pop-up workflow.
Instead, when users upload files, they are now presented with a brief pop-up offering three options: decline AI assistance, view AI organization suggestions, or customize the AI-generated organization. If users choose to customize, all subsequent actions will occur within the AI chat box interface. This conversational flow allows users to interact with the assistant by selecting from predefined prompt keys or typing their own commands. Previous interactions and file states are preserved within the chat history for reference and reversibility.
To provide the user more flexibility, I also designed an expanded view to remove the limitations of a narrow view of the AI chat box. In the expanded view, the entire screen is considered the to be the chat box with the three panel system appearing in a chat bubble.
Since the side-by-side comparison within the pop-up was popular among users during testing, my initial design prototype was selected to be part of the final design. As a team, we collaborated to refine the visual details and ensure the component aligned with the company’s overall branding, maintaining a consistent look and feel across the product.
These are the themes that consistently emerged throughout user testing and are key insights we aim to highlight for future works.
