Important Notes
Key tips and common mechanisms when using your COCO AI employee.
Initial Setup (Required for First Use)
After adding your AI employee to a group chat, don't start asking questions right away — complete these setup steps first:
Step 1: Enable Group Chat Whitelist
Send the following message in the group by @mentioning your AI employee:
@AI-Employee Enable this group's whitelist
The AI employee will confirm that the group is activated. After that, all group members can interact normally.
Step 2: Grant Private Chat Access for Members (Optional)
If you want a specific member to be able to chat 1-on-1 with the AI employee, send in the group:
@AI-Employee Grant @member private chat access
Note: @mentioning the AI employee in group chats is not restricted by the whitelist — all group members can use it. The whitelist only controls private chat permissions.
Step 3: Choose Group Chat Mode (Optional)
| Mode | Trigger | Best For | How to Enable |
|---|---|---|---|
| Standard Mode (Default) | Must @mention AI employee to get a reply | Large groups, daily discussions | Enabled by default, no action needed |
| Smart Group Chat Mode | AI automatically decides whether to respond | Small project groups (<10 people) | @AI-Employee and say "Enable smart group chat mode" |
To enable smart group chat mode:
@AI-Employee Enable smart group chat mode
To disable:
@AI-Employee Disable smart group chat mode
Note: In smart group chat mode, the AI reads all messages in the group to determine whether a response is needed. This increases token consumption. We recommend enabling it only in core working groups.
Send Your First Task
After completing the initial setup, @mention your AI employee in a group chat or send a direct message in private chat:
@AI-Employee Hello! Please introduce yourself and tell me what you can help me with.
The AI employee will respond based on your plan's capabilities. Once you confirm everything works, you're ready for daily use.
Interaction Quick Reference
| Scenario | How to Interact | Notes |
|---|---|---|
| Group chat | @AI-Employee name + message content | In default mode, must @mention to trigger a reply |
| Private chat | Send messages directly | Requires private chat whitelist access first |
| Multi-person group | @AI-Employee, reply visible to all members | AI's reply is visible to everyone in the group |
Tip: Messages in group chats that don't @mention the AI employee will not be processed or replied to — it won't interfere with normal team communication.
Long-Term Memory & SOP Storage
AI employees have long-term memory capabilities. They can remember important workflows (SOPs), preference settings, and business rules, and will continue to follow them in future conversations.
How to Store SOPs
Simply tell your AI employee what you want it to remember during a conversation:
Examples:
Please remember this workflow:
1. Check inbox for customer inquiries every morning at 9 AM
2. Organize inquiries into standard format: client name, requirements, contact info
3. Categorize by urgency: A/B/C levels
4. Level A: respond same day, B: within 48 hours, C: within one week
Please remember: Our brand color is #F5C542. All designs must use this color.
Please remember: When replying to client emails, always use this signature:
Best regards,
[Name]
[Company] | [Title]
Memory Management
| Action | How |
|---|---|
| Add memory | Say "Please remember..." or "From now on, always..." |
| View memories | Ask "What rules have I told you?" or "List all my saved SOPs" |
| Modify memory | Say "Change the previous rule about XX to..." |
| Delete memory | Say "Forget the rule about XX" |
Tip: Long-term memories persist within the same AI employee instance, even across new conversations. We recommend storing important business rules, workflows, and formatting requirements in long-term memory.
What to Store in Long-Term Memory
- Workflows (SOPs): Daily procedures, approval processes, reporting formats
- Brand guidelines: Brand colors, fonts, tone of voice, prohibited words
- Business rules: Pricing strategies, discount policies, service terms
- Communication preferences: Email formats, reply templates, language style
- Team info: Organization structure, contacts, role assignments
Other Common Notes
Message Length Limits
| Channel | Per-Message Limit | Long Message Handling |
|---|---|---|
| Telegram | 4,096 characters | Automatically split into multiple messages |
| Lark/Feishu | ~30,000 characters | Content auto-truncated if exceeded |
| Microsoft Teams | ~28 KB | Automatically split into multiple messages |
| Zalo (Official) | 2,000 characters | Automatically split into multiple messages |
| Zalo Personal (Unofficial) | 2,000 characters | Automatically split into multiple messages |
| Discord | 2,000 characters | Automatically split into multiple messages |
File Handling
AI employees can process the following file types (requires im:resource permission):
- Documents: PDF, Word, Excel, PPT
- Images: PNG, JPG, GIF
- Text: TXT, CSV, JSON, Markdown
Note: File size should not exceed 20MB. Large files may take longer to process.
Response Times
| Scenario | Expected Response Time |
|---|---|
| Simple Q&A | 3-10 seconds |
| File processing | 10-30 seconds |
| Complex analysis | 30-60 seconds |
| Long document generation | 1-3 minutes |
Tip: If your AI employee hasn't responded for over 2 minutes, try resending your message.
Security & Privacy
- Your AI employee will not share your conversation data with other users
- Each AI employee instance's data is isolated
- Avoid sending sensitive passwords or keys in conversations
- Enterprise plans support custom data retention policies