Skip to content

Temporary

Completed Tasks

Revamped Minio Document Playground UI
Implemented loader animation when calling tools in chat
Revamped UI Stepper in Procedures
Resolved Task integration issue in BJB environment
Fixed Document Analyzer bug in BJB
Resolved local procedure Card bug
Fixed language extraction issue in Website Extraction
Resolved issue: Fact not displayed on website
Fixed: SST cannot pause

This Week Next

Advance Prompt Support in Minio (Ensuring fact extraction results are structured and logically grouped in the answer) → need reprod (kemungkinan limit token)
Detail
As a platform user, ​I want the AI to extract and present facts from documents in a well-structured and logically grouped format, ​So that the answers are coherent, readable, and easier to interpret in the context of knowledge retrieval.
Acceptance Criteria:
Facts retrieved from Minio resources are grouped by relevance, theme, or section.
Redundant or disjointed outputs are minimized.
Users can review structured responses in a consistent format.

Download Document Feature in Minio → Fahri -< 1/4 day
Detail
As a platform user, ​I want the AI to extract and present facts from documents in a well-structured and logically grouped format, ​So that the answers are coherent, readable, and easier to interpret in the context of knowledge retrieval.
Acceptance Criteria:
Facts retrieved from Minio resources are grouped by relevance, theme, or section.
Redundant or disjointed outputs are minimized.
Users can review structured responses in a consistent format.

Collaboration Mode → need dosvuss
need reproduce Enable multiple sessions for the same user account when logged in across devices


Performance Optimization

Investigate and improve latency:
Tool - Knowledge (Minio) - Investigate and reduce latency for the following tools, currently experiencing delays of up to 60 seconds. Target performance is under 15 second - Fahri → 1 day
Tool - OSINT (Minio) - Investigate and reduce latency for the following tools, currently experiencing delays of up to 50 seconds. Target performance is under 15 second - Fahri → 1 day QA →Bechmark → Rosma → 1 day

Known Bugs

OSINT Tool fails when configured with multiple tools → Fahri → 1/4 Day
Fails to function properly when configured alongside multiple tools (e.g., Knowledge, Metadata, Data Connection).
Issue: The system consistently defaults to selecting Knowledge Tool, preventing proper multi-tool execution.

Enhancements

Add Retry and Stop options during AI chat response → Fahri → 1 day Enable Retry and Stop options during AI-generated chat responses
Detail
As a user, ​I want to have the ability to retry or stop the AI's response while it is generating an answer, ​So that I can quickly correct errors, adjust input, or prevent unnecessary waiting when the response is not relevant or taking too long.
Acceptance Criteria:
A Stop button is available while the AI is typing, allowing the user to halt the response immediately.
A Retry button is available after stopping or when the response finishes, allowing the user to resend the prompt.
Visual feedback (e.g., spinner, loader) indicates when the AI is processing.
Stopping and retrying should not result in system errors or context loss.
Retry Flow
User Prompt Submission
User submits a prompt to the AI.
AI Begins Generating Response
System displays a loading animation or typing indicator.
During this phase, Stop button is available.
User Clicks “Retry”
Can occur after the response is completed or after clicking Stop.
System Action on Retry
Clears the previous AI response.
Re-sends the exact same prompt to the AI engine.
Triggers a fresh response generation.
Maintains context (unless intentionally reset by user).
AI Generates New Response
New AI output is displayed.
Loading indicator appears again, along with the Stop button.
System Logs Retry
Retry event is logged for analytics (optional).
Flow & Condition
Condition
Expected Behavior
Response is stopped midway
Retry re-sends the full original prompt.
Response completes but is unsatisfactory
Retry sends the same prompt again with a new generation cycle.
Network failure during generation
Retry attempts re-generation once network is stable.
User modifies prompt after stopping
Retry is disabled or replaced with "Send" (not the original prompt anymore).
Multiple retries in a row
Each retry generates a new and unique response unless caching is active.
Retry on system timeout or failure
Retry reinitializes the response logic and session context (if needed).
Retry when context is missing or lost
System displays warning or auto-restores the last available context.
There are no rows in this table

Gemini -BE Support Gemini 2 -> not started -BE Object Breaking -> not started


Want to print your doc?
This is not the way.
Try clicking the ··· in the right corner or using a keyboard shortcut (
CtrlP
) instead.