/voice

Voice Feedback Widget & Analysis

Purpose: Use this file when the task is in-app feedback capture, feedback categorization, sentiment analysis, or user-response templates.

2026 widget landscape. Hotjar, Zonka, Qualaroo, Sprig, FullStory, LogRocket, and dozens of others ship in-app NPS / feedback widgets with ~2-minute setup and AI categorisation built in. The 2026 differentiator is no longer "do you have an AI categoriser" — it is whose codebook the categoriser uses; pipe widget submissions through the team-curated taxonomy and accuracy discipline in thematic-coding.md before they drive routing decisions. The keyword tables below are the fallback for low-volume / sample-of-one feedback, not the production classifier — for any volume worth analysing, use a tuned LLM coder against the same locked codebook.

Tool name changes (2023→): Apptentive (formerly a standalone mobile feedback platform) was acquired by Alchemer in January 2023 and is now branded Alchemer Mobile (https://www.alchemer.com/apptentive-is-now-alchemer-mobile/). Do not reference "Apptentive" as an independent product — use Alchemer Mobile.

Contents:

• Widget design and event contract
• Feedback categories and sentiment rules
• Minimal analysis logic
• Feedback analysis report format
• Close-the-loop response templates

In-App Feedback Widget

Recommended feedback types:

Minimal submission contract:

interface FeedbackSubmission {
  type: 'bug' | 'feature' | 'improvement' | 'praise' | 'other';
  message: string;
  page: string;
  userId?: string;
  screenshot?: string;
}

trackEvent('feedback_submitted', {
  type,
  message_length: message.length,
  page: window.location.pathname
});

Feedback Categorization Framework

Sentiment Classification

Minimal Analysis Logic

interface AnalyzedFeedback {
  original: string;
  sentiment: 'positive' | 'neutral' | 'negative';
  sentimentScore: number;
  categories: string[];
  keywords: string[];
  actionable: boolean;
}

const positiveKeywords = ['convenient', 'good', 'helpful', 'great', 'thanks'];
const negativeKeywords = ['problem', 'slow', 'error', 'confusing', 'hard'];

Rules:

• mark feedback as actionable when it maps to feature, bug, or usability
• allow multi-label categorization
• prefer human review for low-confidence sentiment or sarcasm

Feedback Analysis Report: [Period]

## Feedback Analysis Report: [Period]

### Summary
| Metric | Value | vs Previous Period |
|--------|-------|-------------------|
| Total Feedback | [N] | [+/-X%] |
| NPS Score | [X] | [+/-X points] |
| Positive Sentiment | [X%] | [+/-X%] |
| Negative Sentiment | [X%] | [+/-X%] |

### Category Breakdown
| Category | Count | % of Total | Trend |
|----------|-------|------------|-------|
| Feature Requests | [N] | [X%] | Up/Down/Flat |
| Bug Reports | [N] | [X%] | Up/Down/Flat |
| Usability Issues | [N] | [X%] | Up/Down/Flat |
| Praise | [N] | [X%] | Up/Down/Flat |
| Other | [N] | [X%] | Up/Down/Flat |

### Top Issues
| Rank | Issue | Count | Impact | Recommendation |
|------|-------|-------|--------|----------------|
| 1 | [Issue description] | [N] | [H/M/L] | [Action] |

### Recommended Actions
1. **High Priority:** [Action] - [Expected impact]
2. **Medium Priority:** [Action] - [Expected impact]
3. **Low Priority:** [Action] - [Expected impact]

Close-the-Loop Response Templates

Positive Feedback

Thank you for the feedback. We are glad [specific point] is working well for you.
We will keep investing in that experience.

Feature Request

Thank you for the suggestion. We have logged [feature name] for review alongside similar requests.
We will share an update if it moves forward.

Bug Report

Thank you for reporting this issue. We are sorry for the inconvenience.
We are investigating [issue] and will follow up after we confirm the fix.

Negative Feedback

Thank you for sharing this. We are sorry the experience around [issue] was frustrating.
We are reviewing the feedback seriously and will communicate the next step.