How Lexic Works

The situation

Maya Chen runs a freelance UX design business while writing a historical novel in her spare time. Both are personal projects, but they're completely different mental contexts.

The problem

With everything in one lexicon, the AI kept connecting "character development" (her novel's protagonists) with "user personas" (her UX research). Both involve understanding people—but the novel's 18th-century merchant has nothing to do with her client's SaaS onboarding flow. The knowledge graph became noise.

The solution

Two separate lexicons.

Now when Maya opens her Design Business lexicon, she only sees connections between client work. Her novel research doesn't pollute the signal.

The situation

Jordan is a college junior studying Computer Science with a minor in Philosophy. They're also president of the campus environmental club and planning a study abroad semester.

The problem

Course notes, club planning, and travel research all landed in one lexicon. The AI started connecting "environmental ethics" (Philosophy 301) with "campus sustainability initiatives" (club work)—which seems useful until Jordan realized their essay citations were getting tangled with club meeting notes. Their professor doesn't need to see that the "Smith 2019" reference also appears in notes about recycling bin placement.

The solution

Multiple lexicons for distinct life domains.

Why not just use workspaces?

Jordan considered putting everything in one lexicon with workspaces for each area. But the AI would still try to connect everything. When they're writing a Philosophy paper on environmental ethics, they don't want the AI surfacing notes from last week's club meeting about T-shirt orders.

The situation

A 5-person product squad needs to share strategy with leadership but keep sprint details internal.

The problem

They can't share "the workspace" because some content is sensitive.

The solution

Multiple canvases with different sharing.

Workspace: Q1 Planning
├── Canvas: Strategy (shared with leadership)
├── Canvas: Sprint Notes (squad only)
└── Canvas: Retrospectives (squad only)

The PM invites leadership to the Strategy canvas. Leadership can see those notes, the entities extracted from them, and connections within that canvas. But they can't see Sprint Notes—even though it's in the same workspace.

The situation

A researcher starts with a simple setup—one lexicon, three canvases. Six months later, they have 40 canvases across multiple research projects.

The problem

Finding the right canvas takes forever. The sidebar is a wall of text.

The solution

Add workspaces to group related canvases.

Lexicon: Research
├── Canvas: Project Alpha Interviews
├── Canvas: Project Alpha Synthesis
├── Canvas: Project Beta Survey
├── Canvas: Project Beta Analysis
├── Canvas: Literature Review
├── Canvas: Methods Notes
└── ... (34 more canvases)

Lexicon: Research
├── Workspace: Project Alpha
│   ├── Canvas: Interviews
│   └── Canvas: Synthesis
├── Workspace: Project Beta
│   ├── Canvas: Survey
│   └── Canvas: Analysis
└── Workspace: Methods
    ├── Canvas: Literature Review
    └── Canvas: Methods Notes

Same content. Same AI connections. Just easier to navigate.

The situation

A consultant maintains their own knowledge base but needs to share deliverables with clients.

The problem

They can't give clients access to their internal analysis.

The solution

Separate canvases for internal vs. external work.

Lexicon: Consulting
├── Workspace: Client A
│   ├── Canvas: Internal Analysis (private)
│   └── Canvas: Deliverables (shared with client)
└── Workspace: Client B
    ├── Canvas: Internal Analysis (private)
    └── Canvas: Deliverables (shared with client)

Clients only see their deliverables canvas. The consultant's working notes stay private. If the same entity (like "Budget Constraints") appears in both canvases, the client only sees context from notes they have access to.

The situation

A 200-person company has Marketing, Engineering, and Product teams. Each needs their own space, but they also need shared context.

The solution

Multiple team lexicons with org-shared entities.

Organization: Acme Corp
├── Lexicon: Marketing (brand guidelines, campaigns)
├── Lexicon: Engineering (architecture, incidents)
└── Lexicon: Product (roadmaps, research)

When Product creates a note mentioning "Project Mercury," they can mark that entity as org-shared. Now when Engineering writes implementation notes, they see "Project Mercury" as a recognized entity—with context that this is a real initiative.

But Engineering can't see Product's actual notes. They just know the entity exists.