Oprill / Enterprise graph
See how knowledge connects.
Documents mean more when you can follow their relationships to people, projects and decisions.
Product concept · Interactive website preview
contextReview notesBoundaries
Follow a question across connected records.
A launch checklist may depend on a decision recorded elsewhere. A graph can make the path between those records explicit and inspectable.
Explore relationships
The nodes and edges are synthetic; no organizational data is indexed.
The checklist change traces back to a specific review.
Illustrative exampleConnect the content with the work it describes.
A document title alone rarely explains why it matters. Project membership, ownership and review state add context to a retrieved passage.
Explore depth
An inferred relationship should retain evidence and uncertainty rather than appear as an unquestioned fact.
The source belongs to a recognizable project context.
Illustrative examplePreserve the boundary on every relationship.
A person may be able to see a project without being allowed to read every associated document. Consider titles and edge labels as well as full content.
Explore access
A graph connection must not become an implicit access grant.
The explanation stops before revealing restricted information.
Illustrative exampleSelect a view for the task at hand.
A new teammate and a readiness reviewer need different slices of the same project. The current question gives the context a useful focus.
Explore relevance
These are fixed conceptual views, not personalized profiles.
The review view emphasizes open decisions and dependencies.
Illustrative exampleCarry the relationship evidence into the next step.
An agent should receive both the selected fact and the path that explains it. That helps a reviewer understand why the information was used.
Explore workflows
No agent runs from this page; the diagram explains the handoff.
The draft retains the relationship behind its conclusion.
Illustrative exampleEnterprise graph in context
Trace the decision behind a changed launch checklist.
Interactive illustration with synthetic content. No AI model or company system is connected.
Project Atlas ↔ Launch team ↔ Decision owner
The decision has an owner who can explain its scope.
An organizational relationship is not an access grant. A graph must respect the permissions of the underlying source.
Checklist ↔ Review decision ↔ Dependency note
The checklist change traces back to a dependency review.
Relationships should have provenance. A manually asserted link and a link inferred from text deserve different levels of confidence.
Visible: team checklist / Restricted: private review
The graph explanation stops at the access boundary.
This diagram is synthetic and browser-local. It does not index company content or infer relationships about real people.
Knowledge keeps its context
Different sources. Distinct responsibilities.
Original source examples for enterprise graph. These symbols represent information types, not connected applications.
Questions about enterprise graph
Does a relationship grant access?
No. A connection between a project and a document does not authorize a person to read the document or sensitive metadata.
How should an inferred link be presented?
Label its uncertainty and retain the evidence behind it. Do not present a plausible relationship as a confirmed organizational fact.
Understand the concept. Keep the decision.
These original examples explain a possible approach to enterprise graph. Oprill’s website does not run AI agents, connect to company data or carry out external actions. Your selections stay in this page and reset when you reload.