Setting Up AI
Setup steps
Section titled “Setup steps”- Go to Settings (in the sidebar)
- Scroll down to the AI Assistant section
- Toggle Enable AI Assistant to ON
- Select a Provider from the dropdown:
- OpenAI
- Anthropic
- Google (Gemini)
- Mistral
- OpenRouter
- Select a Model from the dropdown (options depend on provider):
| Provider | Available models |
|---|---|
| OpenAI | GPT-4o, GPT-4o Mini, GPT-4.1, GPT-4.1 Mini |
| Anthropic | Claude Sonnet 4.6, Claude Haiku 4.5 |
| Gemini 2.5 Pro, Gemini 2.5 Flash | |
| Mistral | Mistral Large, Mistral Medium |
| OpenRouter | GPT-4o, Claude Sonnet 4.6, Llama 4 Maverick, DeepSeek R1, Gemini 2.5 Pro |
Select Other… at the bottom to enter a custom model ID.
- Paste your API Key (see Getting an API Key)
- Click Test Connection to verify everything works
- Green message: “Connection successful” — you’re good
- Red message: “Connection failed: [error]” — check your key and settings
- Click Save
After setup
Section titled “After setup”Once saved:
- A teal floating button appears on every page for all team members
- The API key status shows a green dot: “API key is set and encrypted”
- Only the last 4 characters of the key are shown (e.g.,
sk-****abc1)
What non-admin users see
Section titled “What non-admin users see”Users with the Manager or User role see a read-only status in Settings:
- If enabled: “AI Assistant is enabled. Use the floating AI button on any page.”
- If not enabled: “AI Assistant is not enabled. Contact your admin to set it up.”
They cannot see or change the provider, model, or API key.
Switching providers
Section titled “Switching providers”You can change providers at any time:
- Go to Settings > AI Assistant
- Change the Provider dropdown (the model list updates automatically)
- Select a new model
- Enter the new provider’s API key (if different)
- Click Test Connection to verify
- Click Save
Troubleshooting
Section titled “Troubleshooting”| Problem | Solution |
|---|---|
| ”Connection failed: Invalid API key” | Verify the API key is correct. Try regenerating it at your provider. |
| ”Connection failed: Model not found” | Check the model name. Use the dropdown instead of “Other…” if unsure. |
| Team can’t see the AI button | Make sure the Enable toggle is ON and you clicked Save. |
| AI gives poor answers | Try a more capable model (e.g., GPT-4o instead of GPT-4o Mini). |
| ”AI assistant is not enabled” error in chat | The admin hasn’t enabled it yet, or settings weren’t saved. |