Getting started for admins
This guide outlines the steps required to set up the FirstSpirit AI Suite for your team. Each step links to the relevant configuration page.
Connecting a model and vendor
Before your team can use the FirstSpirit AI Suite, you have to connect a model and vendor. You can find all instructions under Connecting a model and vendor.
Activating and configuring assistants
The activation happens on two levels:
- Server-level configuration (local FirstSpirit server only): Activate and configure the AI Suite at the server-level via the ServerManager. For full instructions see Server configuration.
- Project-level configuration: Activate and configure individual assistants for each project via the project components. This step applies to both local server and cloud environments. For full instructions see Project configuration.
Tailoring prompts to your projects
Once the assistants are active, prompts can be configured to reflect your project requirements and brand voice. You can adjust prompts in the ContentCreator or, as a server admin, in the ServerManager.
Which prompts you can edit depends on your role:
| Prompting level | Created and edited by | Scope |
|---|---|---|
| System prompts | Server admin only | All projects, defined assistants |
| Global prompts | Server admin only | All projects |
| Shared prompts | Server admin or project admin | All projects |
| Project prompts | Project admin | This project only |
To give editors a consistent set of reusable tasks, custom operations can be edited, created and made available per assistant. You can find detailed instructions under Prompt administration.
Optional: Connecting tools (API, MCP)
MCP servers and the AI API can be configured to give the AI Suite access to external data sources and tools. This allows operations to draw on information from connected systems, such as a product database or a wiki without requiring editors to switch context.
For more information see AI-native work.