Open Prompt in Deep Research
What's New:
You can now use Open Prompt, Feedier's AI-powered conversational assistant, directly within Deep Research. Ask questions about your feedback data in plain language and get instant, context-aware answers without leaving your analysis.
Deep Research is also available now in Dashboard
How It Works:
Ask anything: Type a question in natural language and receive a streaming AI response in real time.
Slash commands: Use commands like /global-summary, /insight-summary, or /create-action-plan to trigger specific analyses scoped to your current Deep Research filters.
Rich responses: Click on FQL filters, insight badges, or action plan cards directly within the AI's response to explore further.
Persistent history: Your conversations are saved automatically, so you can pick up where you left off.
Hidden & Sensitive Attributes for Admins
What's New:
Admins can now view hidden and sensitive attributes on feedback items. Previously, these attributes were completely inaccessible. Authorized users can now reveal them when needed, with full audit logging for transparency.
How It Works:
Reveal attributes: Click the "View Hidden Attributes" button in the feedback attribute section (visible to admins and staff only).
Confirmation required: A confirmation modal warns you that this action will be logged and visible to all team admins.
Auto-hide: Attributes are automatically hidden again after 60 seconds or when you navigate to a different page.
Audit trail: Every access is logged with user and timestamp, visible to all admins in the team.
In next iterations we will add a history feature of the feedback were admins will access to who saw the sensitive information
NLP Attributes Rework
What's New:
Feedier now clearly distinguishes NLP attributes, used for automatic text analysis, from regular custom attributes. You can see which attributes are managed by the NLP engine, create new ones, and control how sensitive data is handled in verbatims.
How It Works:
New NLP type: Attributes used for automatic text extraction are now labeled as NLP type, making them easy to identify in your attribute settings.
Create or convert: You can create new NLP attributes manually or change an existing attribute's type to/from NLP -> description is required to help the llm understand the objective.
Automatic anonymization: Mark an NLP attribute as sensitive and its values are automatically anonymized (replaced with ***) in Live Feed and Text Analysis views.
Deletion control: If you delete an NLP attribute, it will not be recreated automatically, giving you full control over your attribute setup.
CSV Import Error Details
What's New:
When a CSV import fails due to a column name issue, the error message now tells you exactly which column caused the problem, no more guessing.
How It Works:
Clear error messages: If a column name exceeds the maximum allowed length, the error message includes the specific column name so you can fix it immediately.
Faster troubleshooting: No need to check each column manually. The platform points you directly to the issue.