Docs menu
Who can do what
Platform roles, organization roles and team roles, and what each one controls.
In short
Access works on three levels: your platform role, your organization role and your team roles. Your organization role opens features: the parts of Northstar you can use. Your team roles open resources: the datasets, KPIs, figures, dashboards and collections you can use in those features. Most actions need both: the feature, and the right on the resource.
Three levels
| Level | Decides | Set by |
|---|---|---|
| Platform role | Whether you administer the whole platform | A platform administrator |
| Organization role | Which features you can use in an organization, and which actions | An organization administrator |
| Team role | Which resources you can use, and which actions on each | A team administrator |
An organization is the space your company works in. Its datasets, KPIs, dashboards, teams and roles belong to it. If you belong to several organizations, the organization switcher at the top of the sidebar changes the active one.
Platform role
- User: the default. What you can do depends on your organization and team roles.
- Administrator: manages users, organizations and platform settings, and can see and use every resource, in every organization.
Platform administrators also choose which AI features are on, for the whole platform. See AI assistants.
Features and resources
| Features | Resources | |
|---|---|---|
| Examples | Datasets, KPIs, Chat, SQL Executor, Lineage | The dataset "Orders 2026", the dashboard "Sales review" |
| Granted by | Your organization role | Your team roles |
| Answers | "Can I open the Datasets page? Can I create a KPI?" | "Which datasets do I see there? Can I explore this one?" |
Most actions need both. To open a dataset, your organization role needs the Datasets feature, and one of your teams needs the View right on that dataset. To ask Chat about it, you also need the Chat feature and the Ask AI right.
Organization roles group features in three sections:
- Organization management: roles, members, invitations, teams and developer accounts;
- Data management: storage systems, sources, ingestions, datasets, transformations, documents, KPIs, lineage and workflows;
- Analysis tools: Smart Cockpit, Chat, figures, SQL Executor and dashboards.
Default roles
Each organization starts with three roles:
| Role | In short |
|---|---|
| Superadmin | Every feature and every action. |
| Admin | Manages the organization (roles, members, invitations, teams, developer accounts) and most data features. |
| Member | Reads the data features; creates KPIs and figures; uses the Smart Cockpit, Chat and the SQL Executor. |
Organization administrators can create other roles. A new role starts with the Member permissions and can then be adjusted, action by action.
"I can't see it": a checklist
- Is the feature in your organization role? If a page is missing from the sidebar, your role does not include that feature.
- Is the resource shared with one of your teams? A new dataset is only visible to platform administrators until a team is given access to it. The same goes for a silver or gold dataset whose inputs have no team in common.
- Does your team have the right action? Seeing a dataset does not mean exploring it or asking AI about it: each is a separate right.
- Are you in the right organization? Check the organization switcher.
- Is the AI feature on? A platform administrator may have turned it off.
- Is the dataset deprecated? Chat does not answer from deprecated datasets.
If all of these look right, ask the owner of the resource or an organization administrator.
Developer accounts and API keys
Data pipelines and integrations use developer accounts, also called integration accounts: machine accounts managed on the Developers page.
- A developer account has a platform role, an organization role and teams, like a person. It sees what those give it, and nothing more.
- It authenticates with API keys, which belong to one organization and are shown once.
- Give a developer account the platform administrator role only for trusted integrations: it would then see everything.
The Developers section explains how to create these accounts, manage their keys and use the Python SDK.
FAQ
What is the difference between an organization administrator and a platform administrator? An organization administrator manages one organization: its roles, members and teams. A platform administrator manages the whole platform and sees every resource.
I created a dataset. Why can't my colleague see it? It is not shared with any of their teams yet. See teams and resources.
Who can change my organization role? An organization administrator, on the organization page.