Insight Cards
Reinventing the medium for user to get more out of insights
(Domain)
Analytics
(Year)
2024
(Services)
Product Design
About
The Product
Feedback Cards unify feedback, summaries, predictions, and metadata into one source of truth.
The product helps companies analyze and act on the Voice of the Customer by aggregating feedback from channels like app stores, support tools, communities, and call transcripts. Each piece of feedback is structured into a Feedback Card, The single source of truth containing raw feedback (or transcripts), system-generated summaries, predictions (keywords/reasons), and metadata for filtering and context.
Despite being central to the experience, Feedback Cards grew cluttered and outdated, making insight extraction inefficient and pushing users toward external links. After the Querybuilder Redesign, they became the next priority for improvement.
Context
& Problem
As the product scaled, the feedback card became the core surface for customer voice—housing feedback, predictions, summaries, and metadata. But its design and information model failed to support this role, creating friction, inconsistency, and overload that hurt adoption and retention.
Key issues went beyond visual polish:
Scalability and consistency issues: The rigid design struggled to support emerging data types (audio, AI summaries, evolving taxonomy), while inconsistent layouts across feedback types weakened trust and usability.
Information hierarchy gaps: Snippets, predictions, and metadata weren’t prioritized effectively, forcing users to parse through clutter before finding value, slowing analysis and diminishing confidence in insights.
Workflow inefficiency: Overloaded layouts and limited readability disrupted user flow, often pushing users toward external tools for faster, clearer extraction and comparison.
These compounded into product risks: reduced trust, underutilized features, and slower decisions. The redesign rebuilt clarity, scalability, and trust at the platform’s core.
Goals
The redesign aimed to reposition the Feedback Card as the core unit of insight discovery by fixing hierarchy, readability, and interaction gaps. The objectives were:
Enhance clarity and focus: Redefine the information hierarchy to surface predictions, summaries, and metadata upfront—guiding users toward the most actionable insights with improved readability and scanability.
Streamline interactions and context: Simplify actions by prioritizing high-value controls, improving transparency of filter matches and metadata to make relevance instantly clear.
Design for consistency and scalability: Unify feedback formats across text, audio, and other types, and build a modular system flexible enough to support evolving data types and future AI-driven features.
The goal: evolve the card from a noisy container into a focused, scalable surface that builds trust, improves retention, and anchors feedback exploration.
Initial Research
We consolidated customer tickets, VoC calls, and community feedback, which consistently flagged two pain points: the list view was hard to scan beyond one or two records, and snippets often lacked context when filters were applied.
To validate, I combined qualitative and behavioral research:
Session reviews showed many users skipped the list view entirely, diving straight into full record detail and manually copying snippets into notes.
Event data analysis revealed the most common actions were expanding records, opening external links, and copying links, while snippet navigation and collections were rarely used, signaling low value in audio-related controls.
This confirmed the issue wasn’t about adding more actions, users needed better readability, clearer hierarchy, and stronger context.
For benchmarks, I analyzed competing platforms. Patterns were consistent: show only 2–3 lines of relevant context, surface minimal but essential metadata, and sharpen hierarchy with summaries and filter matches above secondary details.
Together, the findings pointed to a clear redesign goal: strip back clutter, emphasize brevity, and surface context intelligently, positioning the feedback card as a focused, high-signal insight artifact.
Lo-Fi Designs
I iterated through three wireframe variations to rethink the card layout:
Tabular Approach: Split space evenly (50/50) between snippets and metadata, adjusted depending on context (full page vs. side-pane).
Condensed Feedback: Prioritized a 3–4 line snippet with metadata below, following a clean top-to-bottom flow for easier scanning.
Combined Variant: A middle ground with 2–3 snippet lines and select metadata.
Outcome: Internal reviews leaned toward the Condensed Feedback format, which felt the most natural for quick scanning. Feedback also emphasized the need for gutter spacing on larger screens for readability. As part of this shift, we also renamed “Search” to Feed, better reflecting its role as a live stream of customer voices.
Hi-Fi Designs
The high-fidelity prototypes translated these insights into a polished redesign:
Feed-based layout: Users saw a cleaner feed of records, optimized for skimming.
Condensed cards: Each record preview focused on the most relevant snippet, with secondary metadata tucked neatly below.
Detail view improvements: Cleaner design that prioritized raw feedback, summaries, and predictions without overwhelming users.
Integration with “Chat with Wisdom”: Allowed users to query AI-powered insights directly from the feed, deepening the utility of cards.
We rolled out the Beta version internally and to 10 customer orgs. Feedback was overwhelmingly positive: users felt the design provided immediate context, allowed faster scanning, and gave more confidence in deciding which records to open.
Outcome
Improved insight collection: Users skimmed more records per session and reported it was easier to decide what to explore vs. skip.
User retention: Reliance on external links dropped from ~58% to ~15%, showing users preferred consuming feedback within the platform.
Business impact: The redesign rebuilt trust with long-standing users by addressing top frustrations, helping retain accounts and improving daily engagement.
Takeaway
What worked: A leaner, decluttered design improved readability and focus.
What was missing: Audio cards saw limited change, with demand for video, AI chapters, and richer formats scoped for future work.
Key learnings: I built the redesign case by synthesizing user pain from tickets, VOC, and analytics, sharpening my ability to balance urgent fixes with scalable foundations.
✨ The Feedback Card Redesign proved that clarity drives adoption, turning cluttered cards into focused insight units, restoring trust, and setting the stage for richer, future-ready formats.
Insight Cards
Reinventing the medium for user to get more out of insights
(Domain)
Analytics
(Year)
2024
(Services)
Product Design
About
The Product
Feedback Cards unify feedback, summaries, predictions, and metadata into one source of truth.
The product helps companies analyze and act on the Voice of the Customer by aggregating feedback from channels like app stores, support tools, communities, and call transcripts. Each piece of feedback is structured into a Feedback Card, The single source of truth containing raw feedback (or transcripts), system-generated summaries, predictions (keywords/reasons), and metadata for filtering and context.
Despite being central to the experience, Feedback Cards grew cluttered and outdated, making insight extraction inefficient and pushing users toward external links. After the Querybuilder Redesign, they became the next priority for improvement.
Context
& Problem
As the product scaled, the feedback card became the core surface for customer voice—housing feedback, predictions, summaries, and metadata. But its design and information model failed to support this role, creating friction, inconsistency, and overload that hurt adoption and retention.
Key issues went beyond visual polish:
Scalability and consistency issues: The rigid design struggled to support emerging data types (audio, AI summaries, evolving taxonomy), while inconsistent layouts across feedback types weakened trust and usability.
Information hierarchy gaps: Snippets, predictions, and metadata weren’t prioritized effectively, forcing users to parse through clutter before finding value, slowing analysis and diminishing confidence in insights.
Workflow inefficiency: Overloaded layouts and limited readability disrupted user flow, often pushing users toward external tools for faster, clearer extraction and comparison.
These compounded into product risks: reduced trust, underutilized features, and slower decisions. The redesign rebuilt clarity, scalability, and trust at the platform’s core.
Goals
The redesign aimed to reposition the Feedback Card as the core unit of insight discovery by fixing hierarchy, readability, and interaction gaps. The objectives were:
Enhance clarity and focus: Redefine the information hierarchy to surface predictions, summaries, and metadata upfront—guiding users toward the most actionable insights with improved readability and scanability.
Streamline interactions and context: Simplify actions by prioritizing high-value controls, improving transparency of filter matches and metadata to make relevance instantly clear.
Design for consistency and scalability: Unify feedback formats across text, audio, and other types, and build a modular system flexible enough to support evolving data types and future AI-driven features.
The goal: evolve the card from a noisy container into a focused, scalable surface that builds trust, improves retention, and anchors feedback exploration.
Initial Research
We consolidated customer tickets, VoC calls, and community feedback, which consistently flagged two pain points: the list view was hard to scan beyond one or two records, and snippets often lacked context when filters were applied.
To validate, I combined qualitative and behavioral research:
Session reviews showed many users skipped the list view entirely, diving straight into full record detail and manually copying snippets into notes.
Event data analysis revealed the most common actions were expanding records, opening external links, and copying links, while snippet navigation and collections were rarely used, signaling low value in audio-related controls.
This confirmed the issue wasn’t about adding more actions, users needed better readability, clearer hierarchy, and stronger context.
For benchmarks, I analyzed competing platforms. Patterns were consistent: show only 2–3 lines of relevant context, surface minimal but essential metadata, and sharpen hierarchy with summaries and filter matches above secondary details.
Together, the findings pointed to a clear redesign goal: strip back clutter, emphasize brevity, and surface context intelligently, positioning the feedback card as a focused, high-signal insight artifact.
Lo-Fi Designs
I iterated through three wireframe variations to rethink the card layout:
Tabular Approach: Split space evenly (50/50) between snippets and metadata, adjusted depending on context (full page vs. side-pane).
Condensed Feedback: Prioritized a 3–4 line snippet with metadata below, following a clean top-to-bottom flow for easier scanning.
Combined Variant: A middle ground with 2–3 snippet lines and select metadata.
Outcome: Internal reviews leaned toward the Condensed Feedback format, which felt the most natural for quick scanning. Feedback also emphasized the need for gutter spacing on larger screens for readability. As part of this shift, we also renamed “Search” to Feed, better reflecting its role as a live stream of customer voices.
Hi-Fi Designs
The high-fidelity prototypes translated these insights into a polished redesign:
Feed-based layout: Users saw a cleaner feed of records, optimized for skimming.
Condensed cards: Each record preview focused on the most relevant snippet, with secondary metadata tucked neatly below.
Detail view improvements: Cleaner design that prioritized raw feedback, summaries, and predictions without overwhelming users.
Integration with “Chat with Wisdom”: Allowed users to query AI-powered insights directly from the feed, deepening the utility of cards.
We rolled out the Beta version internally and to 10 customer orgs. Feedback was overwhelmingly positive: users felt the design provided immediate context, allowed faster scanning, and gave more confidence in deciding which records to open.
Outcome
Improved insight collection: Users skimmed more records per session and reported it was easier to decide what to explore vs. skip.
User retention: Reliance on external links dropped from ~58% to ~15%, showing users preferred consuming feedback within the platform.
Business impact: The redesign rebuilt trust with long-standing users by addressing top frustrations, helping retain accounts and improving daily engagement.
Takeaway
What worked: A leaner, decluttered design improved readability and focus.
What was missing: Audio cards saw limited change, with demand for video, AI chapters, and richer formats scoped for future work.
Key learnings: I built the redesign case by synthesizing user pain from tickets, VOC, and analytics, sharpening my ability to balance urgent fixes with scalable foundations.
✨ The Feedback Card Redesign proved that clarity drives adoption, turning cluttered cards into focused insight units, restoring trust, and setting the stage for richer, future-ready formats.
Insight Cards
Reinventing the medium for user to get more out of insights
(Domain)
Analytics
(Year)
2024
(Services)
Product Design
About
The Product
Feedback Cards unify feedback, summaries, predictions, and metadata into one source of truth.
The product helps companies analyze and act on the Voice of the Customer by aggregating feedback from channels like app stores, support tools, communities, and call transcripts. Each piece of feedback is structured into a Feedback Card, The single source of truth containing raw feedback (or transcripts), system-generated summaries, predictions (keywords/reasons), and metadata for filtering and context.
Despite being central to the experience, Feedback Cards grew cluttered and outdated, making insight extraction inefficient and pushing users toward external links. After the Querybuilder Redesign, they became the next priority for improvement.
Context
& Problem
As the product scaled, the feedback card became the core surface for customer voice—housing feedback, predictions, summaries, and metadata. But its design and information model failed to support this role, creating friction, inconsistency, and overload that hurt adoption and retention.
Key issues went beyond visual polish:
Scalability and consistency issues: The rigid design struggled to support emerging data types (audio, AI summaries, evolving taxonomy), while inconsistent layouts across feedback types weakened trust and usability.
Information hierarchy gaps: Snippets, predictions, and metadata weren’t prioritized effectively, forcing users to parse through clutter before finding value, slowing analysis and diminishing confidence in insights.
Workflow inefficiency: Overloaded layouts and limited readability disrupted user flow, often pushing users toward external tools for faster, clearer extraction and comparison.
These compounded into product risks: reduced trust, underutilized features, and slower decisions. The redesign rebuilt clarity, scalability, and trust at the platform’s core.
Goals
The redesign aimed to reposition the Feedback Card as the core unit of insight discovery by fixing hierarchy, readability, and interaction gaps. The objectives were:
Enhance clarity and focus: Redefine the information hierarchy to surface predictions, summaries, and metadata upfront—guiding users toward the most actionable insights with improved readability and scanability.
Streamline interactions and context: Simplify actions by prioritizing high-value controls, improving transparency of filter matches and metadata to make relevance instantly clear.
Design for consistency and scalability: Unify feedback formats across text, audio, and other types, and build a modular system flexible enough to support evolving data types and future AI-driven features.
The goal: evolve the card from a noisy container into a focused, scalable surface that builds trust, improves retention, and anchors feedback exploration.
Initial Research
We consolidated customer tickets, VoC calls, and community feedback, which consistently flagged two pain points: the list view was hard to scan beyond one or two records, and snippets often lacked context when filters were applied.
To validate, I combined qualitative and behavioral research:
Session reviews showed many users skipped the list view entirely, diving straight into full record detail and manually copying snippets into notes.
Event data analysis revealed the most common actions were expanding records, opening external links, and copying links, while snippet navigation and collections were rarely used, signaling low value in audio-related controls.
This confirmed the issue wasn’t about adding more actions, users needed better readability, clearer hierarchy, and stronger context.
For benchmarks, I analyzed competing platforms. Patterns were consistent: show only 2–3 lines of relevant context, surface minimal but essential metadata, and sharpen hierarchy with summaries and filter matches above secondary details.
Together, the findings pointed to a clear redesign goal: strip back clutter, emphasize brevity, and surface context intelligently, positioning the feedback card as a focused, high-signal insight artifact.
Lo-Fi Designs
I iterated through three wireframe variations to rethink the card layout:
Tabular Approach: Split space evenly (50/50) between snippets and metadata, adjusted depending on context (full page vs. side-pane).
Condensed Feedback: Prioritized a 3–4 line snippet with metadata below, following a clean top-to-bottom flow for easier scanning.
Combined Variant: A middle ground with 2–3 snippet lines and select metadata.
Outcome: Internal reviews leaned toward the Condensed Feedback format, which felt the most natural for quick scanning. Feedback also emphasized the need for gutter spacing on larger screens for readability. As part of this shift, we also renamed “Search” to Feed, better reflecting its role as a live stream of customer voices.
Hi-Fi Designs
The high-fidelity prototypes translated these insights into a polished redesign:
Feed-based layout: Users saw a cleaner feed of records, optimized for skimming.
Condensed cards: Each record preview focused on the most relevant snippet, with secondary metadata tucked neatly below.
Detail view improvements: Cleaner design that prioritized raw feedback, summaries, and predictions without overwhelming users.
Integration with “Chat with Wisdom”: Allowed users to query AI-powered insights directly from the feed, deepening the utility of cards.
We rolled out the Beta version internally and to 10 customer orgs. Feedback was overwhelmingly positive: users felt the design provided immediate context, allowed faster scanning, and gave more confidence in deciding which records to open.
Outcome
Improved insight collection: Users skimmed more records per session and reported it was easier to decide what to explore vs. skip.
User retention: Reliance on external links dropped from ~58% to ~15%, showing users preferred consuming feedback within the platform.
Business impact: The redesign rebuilt trust with long-standing users by addressing top frustrations, helping retain accounts and improving daily engagement.
Takeaway
What worked: A leaner, decluttered design improved readability and focus.
What was missing: Audio cards saw limited change, with demand for video, AI chapters, and richer formats scoped for future work.
Key learnings: I built the redesign case by synthesizing user pain from tickets, VOC, and analytics, sharpening my ability to balance urgent fixes with scalable foundations.
✨ The Feedback Card Redesign proved that clarity drives adoption, turning cluttered cards into focused insight units, restoring trust, and setting the stage for richer, future-ready formats.