2026.8.31 Product Update
This month, Qiabot improved conversation stability, workspace experience, and App capabilities, including App visitor blocking, web refresh prompts, and App crash report collection.

This month, we optimized Qiabot around conversation stability, workspace experience, and App capabilities.
This update focuses on improving mobile abnormal visitor handling, conversation assignment and closing logic, image and copy experience in the workspace, while also adding web refresh prompts and App crash report collection to make daily support work more stable and smoother.
New Capabilities
- Added visitor blocking in the App
Agents can now handle abnormal visitors directly in the App.
When agents encounter spam, abnormal access, or clearly invalid inquiries during mobile reception, they can block the visitor in the App without switching back to the desktop workspace.
- Added web refresh prompts
This update adds refresh prompts for the web version, making the usage process clearer.
When a page needs to be refreshed or when the current status changes, the system can show a clearer prompt. This helps agents reduce misoperations and judge more quickly whether the current page needs to be reloaded.
- Added App crash report collection
To support ongoing App stability improvements, this update adds App crash report collection.
When the App crashes unexpectedly, the system can collect related error information. This helps the technical team locate issues faster and continue improving the mobile experience.
Conversation Stability Improvements
- Optimized conversation assignment logic
We optimized the conversation assignment logic to reduce assignment exceptions.
For teams with multiple agents, accurate assignment directly affects reception efficiency. After this optimization, the process of receiving, assigning, and handling conversations is more stable.
- Optimized conversation closing logic
We optimized the conversation closing logic to reduce cases where conversations were not ended on time.
With more accurate conversation status, agents can more clearly tell which conversations have been handled and which ones still need follow-up.
- Fixed cases where conversations could not be ended manually
We fixed an issue where agents were unable to manually end conversations in some scenarios.
After the fix, agents can more reliably close conversations after handling visitor questions, keeping the workspace conversation list cleaner.
- Fixed the "Pending First Response" status not disappearing after an agent replied
Previously, in some cases, the "Pending First Response" status could remain even after an agent had replied to the visitor.
After this fix, the status now updates more accurately based on agent replies, reducing misjudgment and duplicate handling.
- Increased the initial conversation limit
We increased the initial conversation limit to improve reception capacity.
For teams with higher inquiry volume or concentrated visitor traffic, a higher initial limit helps the system handle more conversations and reduce pressure during peak periods.
Workspace Experience Improvements
- Fixed mouse wheel adjustment issues when viewing conversation images
We fixed mouse wheel adjustment issues when viewing images in conversations.
When agents view screenshots, images, or attachments sent by visitors, they can browse and adjust images more smoothly, improving the efficiency of handling image-based issues.
- Fixed extra spaces when copying content from the conversation box
We fixed an occasional issue where extra spaces appeared when copying content from the conversation box.
With more accurate copied content, agents can more easily forward information, organize records, or copy key details provided by visitors.
- Fixed occasional missing images in automatic messages
We fixed an issue where images in automatic messages occasionally failed to display.
Automatic messages are often used for welcome messages, fixed prompts, and offline reception. More stable image display reduces visitor misunderstanding and keeps automatic message content complete.


