Admin setup checklist
The order to configure an organization in, and why that order.
This is the fastest working path from an empty organization to members answering tickets. Roughly thirty minutes, most of it pasting credentials.
Do it in this order
Integrations before playbooks: a playbook's data-source picker can only offer integrations that already exist. Playbooks before teams matters less, but teams before team-scoped playbooks does.
Buy the seats you need
The trial is one seat, so the Members table has nowhere to put a second person until you have bought seats. Start on the Account page. See Account and billing.
Add at least one AI provider
Nothing investigates without a model. AI Providers takes one credential set per provider, shared organization-wide. Pick one with embedding support if you want similar-ticket search to work. See AI providers.
Add your shared data sources
Integrations is where the organization's databases, logs, metrics and error tracking are configured once for everyone. Use read-only credentials. See Integrations.
Set a spend cap
Before anyone starts running investigations, not after. It takes ten seconds and it's the only backstop against a bad prompt in a loop. See Spending.
Write playbooks
Start with two or three that cover your highest-volume ticket categories. A playbook is what turns a generic investigation into a good one. See Playbooks and the example playbooks.
Give it something to read
Upload runbooks, product notes and exported sheets on the Knowledge page, or connect Notion or Confluence under Integrations. Until one of those is done, admins see a reminder on every portal page.
Create teams, if you need them
Only if some playbooks should be visible to a subset of people. Skip it otherwise. See Teams.
Invite members
Each gets a username, email, password and role. Send them the desktop install guide. See Members.
Verify
Within about fifteen minutes of a member signing in, the Integrations page should show your shared data sources as last-reported running with non-zero tool counts. That is your end-to-end proof that configuration reached a real machine and started there.
What you are administering, exactly
| Page | Who sees it | What it controls |
|---|---|---|
| Organization | Everyone (edit: admins) | Members, teams, spend cap, which data sources are pre-added for members |
| Playbooks | Everyone (edit: admins) | Triage profiles: instructions, routing, tool scope, visibility |
| Integrations | Everyone (edit: admins) | Shared MCP data sources and their credentials |
| AI Providers | Admins only | LLM credentials, organization-wide |
| Audit | Admins only | Append-only trail of every configuration change |
| Account | Owner only | Plan, seats, billing, activated devices |
What you are not administering here
Tickets, investigations, Console threads, usage totals and report schedules all live on the desktop. See Architecture for why the line falls where it does.