Feedback
The feedback module lets you collect opinions, requests, and reports from your users across multiple channels and centralize them in Almirant. AI automatically clusters similar feedback and suggests actions, connecting the voice of your users directly to your ideation and planning process.
Full workflow
Feedback follows a journey from collection to implementation:
- Collection -- Feedback arrives from any of the four supported sources
- Processing -- AI analyzes the feedback, groups similar entries (clustering), and generates suggestions
- Linking -- Feedback is linked to existing or new ideas
- Action -- Ideas with enough feedback are approved and promoted to work items
Feedback sources
Embedded widget
Integrate a widget directly into your product so users can submit feedback without leaving the application. The widget captures the user's message along with contextual metadata (current page, session, etc.).
Telegram bot
Connect a Telegram bot to receive feedback directly from a group or channel. Messages sent to the bot are recorded as feedback entries in Almirant with the source telegram.
Public API
Send feedback programmatically from any external system using the public API:
POST /api/feedback/ingest
The endpoint accepts a JSON payload with the feedback content and optional metadata. This lets you integrate feedback collection into custom forms, chatbots, mobile apps, or any service that can make HTTP requests.
Manual
Create feedback entries manually from the Almirant interface. Useful for recording feedback received via email, phone calls, meetings, or any other non-automated channel.
Feedback listing
The main feedback view shows all received entries with the following information:
| Field | Description |
|---|---|
| Content | Feedback text submitted by the user |
| Source | Origin channel (widget, telegram, api, manual) |
| Date | Date and time received |
| Cluster | Thematic group assigned by AI |
| Linked idea | Idea it is linked to (if applicable) |
AI clustering
When feedback accumulates, AI analyzes the content and automatically groups entries that deal with the same topic or request. This enables:
- Pattern identification -- Detect which feature or improvement multiple users are requesting
- Demand quantification -- See how many users are asking for the same thing
- Prioritization -- Make decisions based on actual feedback volume
Each cluster has a descriptive name generated by AI (for example, "PDF export request" or "Image loading issues") and groups all related feedback entries.
How it works
- Each new feedback entry is analyzed against existing clusters
- If the content is similar to an existing cluster, it is added automatically
- If it doesn't fit any cluster, a new one can be created
- Clusters are updated dynamically as more feedback arrives
AI suggestions
Based on feedback clusters, AI can generate suggestions such as:
- Create an idea from a high-volume cluster
- Link feedback to an existing idea that already addresses the request
- Prioritize ideas that have the most associated feedback
Suggestions appear as recommendations that you can accept or dismiss.
Link feedback to ideas
The most important step is connecting feedback with ideas to create traceability between what users are requesting and what the team decides to build.
Link from feedback
From a feedback entry's detail view, click Link to idea and select an existing idea or create a new one. This establishes a direct relationship between the feedback and the idea.
Link from the idea
From an idea's detail view, you can search and link existing feedback entries. This is useful when you already have a created idea and want to associate the feedback that justifies it.
Unlink feedback
If you link feedback to an idea by mistake, you can unlink the relationship from either the feedback detail or the idea detail. The feedback entry is not deleted -- only the association is removed.
The power of traceability
When feedback is linked to ideas, and ideas are promoted to work items, a complete chain is created:
This chain lets you answer questions like:
- "Why are we building this?" -- Because 15 users requested it (see linked feedback)
- "Which users benefit?" -- Those who submitted feedback on this topic
- "What impact does this task have?" -- It resolves the largest feedback cluster of the month
See Promotion and traceability for more details on the full workflow.
Feedback exposes the following endpoints and MCP tools:
Public API:
| Endpoint | Description |
|---|---|
POST /api/feedback/ingest | Feedback ingestion from external sources |
GET /api/feedback/bootstrap | Gets the initial configuration for the feedback widget |
MCP tools for linking:
| Tool | Description |
|---|---|
link_feedback_to_idea_item | Link a feedback entry to an idea |
unlink_feedback_from_idea_item | Unlink feedback from an idea |
The ingestion endpoint (/api/feedback/ingest) is public and does not require session authentication, which allows it to be integrated into client-side widgets or external services. It accepts a JSON payload with the feedback content and optional metadata such as source and user identifier.
Related: Idea management | Promotion and traceability