Core concepts
The mental model: workspace, context, automations, insights.
Workspace
A workspace is your team's home in Ion. It contains your members, integrations, docs, automations, and activity. Workspaces are isolated — data never crosses between them.
Context
Context is everything Ion knows about your work: docs, tickets, threads, commits, and the relationships between them. Ion builds a living knowledge graph that updates in real-time as your team works.
Context is what makes Ion's AI useful. When you ask a question, Ion answers from your context — with citations — rather than from a foundation model's training data.
Automations
An automation is a multi-step workflow: when X happens, do Y, then Z. Automations have:
- Triggers — events from your integrations (PR merged, ticket created, message posted).
- Steps — actions Ion takes (filter, transform, call an API, post a message, update a ticket).
- Error paths — what to do when a step fails (retry, skip, notify).
Every automation has a per-step audit log and one-click rollback to the last known good state.
Insights
Insights are dashboards built from the work your team is already doing. Cycle time, throughput, stuck work — all auto-calculated, no manual data entry.