Where enterprise knowledge becomes actionable.
A knowledge platform for big companies, so people can find work that already exists instead of building it twice.
The people who need past work and the people who have it don’t connect
In a big company, the work someone needs usually already exists. They just can’t reach it, so they redo it.
Meet HiveMind — one place to search everything the company knows
HiveMind keeps past projects, the people who worked on them, and related research together in one place. You only see what you have access to.
From Problem to Product
HiveMind helps enterprise teams rediscover past work, connect organizational knowledge, and manage projects across disconnected tools through AI-powered search and secure collaboration.
The animation illustrates the design process behind the solution ← starting with the underlying problem (Why), defining the design approach (How), and culminating in the final product (What).
Find knowledge you didn't know existed
This is what HiveMind is for, so it gets the most space. Someone opens the app, looks for past work, and eventually runs into a file they don’t have access to. Here’s how each screen handles that.
Start a project by describing it
Instead of setting up a project from scratch, you describe what you’re working on and HiveMind builds the first version.
See your tasks without leaving HiveMind
HiveMind isn't only for finding old work — it's also where you track work that's happening right now, even if your team runs it in Jira, Monday, or something else entirely.
Help without the context switch
Rather than centering the product around a chatbot, HiveAI appears only when it can help someone complete the task they’re already working on. It supports research, project creation, and knowledge discovery without becoming another destination to navigate.
Four things kept coming up
This came from secondary research, workplace forums, and talking to people in product and research roles. I also worked through a competitive analysis, a SWOT, and a logic framework matrix to test feasibility and define what success would actually need to look like. I built one persona to represent the pattern, and four findings shaped the rest of the project.
| Capability | Trello | Confluence | Asana | Monday | Jira | HiveMind |
|---|---|---|---|---|---|---|
| AI recommendations | ✕ | ✕ | ✕ | ✕ | ✕ | ✓ |
| Centralized project database | ✕ | ✓ | ✕ | ✓ | ✓ | ✓ |
| Integrated research tools | ✕ | ✕ | ✕ | ✕ | ✓ | ✓ |
| Clearance-based access | ✕ | ✕ | ✕ | ✕ | ✓ | ✓ |
| AI chatbot assistance | ✕ | ✕ | ✕ | ✕ | ✕ | ✓ |
The design system
It was important to root my design in a formal process so I could work as efficiently as possible. Typography and color theory were fundamental in setting the tone to make the product feel serious. When I brought AI into the workflow later, this same design system was what kept everything aligned.
Validated flows
Before any hi-fi work, I mapped the decision logic for each core feature — dashboard, projects, database, chat — in FigJam first. Strategizing the flows before the screens gave me a sense of the most logical path a user would take versus the edge cases, which guided which frames I prioritized designing first.
Low-fi to AI-built
I refined HiveMind through four iterations, each focused on a different goal. Starting with rough concepts to define the experience, I gradually refined the layout, visual system, and interactions before transforming the final designs into a fully functional prototype using AI-assisted development.
I didn’t stop at the designs.
Rather than moving through a linear handoff from design to development, I worked in an iterative loop. Each pass translated designs into code, validated the implementation, refined what didn’t match, uncovered edge cases, and repeated until the product behaved the way it was intended.
Decisions
What I took away
The research and strategy pushed me more than the visual design did. Designing a complex, data-dense enterprise application required balancing information architecture, AI interactions, scalable workflows, and feature prioritization long before polishing the interface. Building HiveMind into a functional front-end reinforced those decisions, exposing interaction details and edge cases I never would have caught in Figma alone.
Where else it fits
From healthcare systems to universities and government agencies, any organization managing complex institutional knowledge faces the same challenge: helping people find what they need without compromising security.
Key learnings
- I became more confident cutting ideas than adding them — Building exposed which features were essential and which weren't at a faster rate.
- AI sped up iteration, but not product thinking — I learned to adapt new AI tools without compromising my design decisions.