Tiffany Lim
Work
About
Contact
The Problem
When too many comments become too much of a good thing
Slickdeals is powered by its community, and popular deals often spark hundreds of comments filled with insights. But that volume came at a cost. Users struggled to find answers to key questions that impacted their decision making. Critical information got buried, and many gave up before getting what they needed. What was meant to build trust and connection had become overwhelming noise, causing frustration and drop-off instead of helping users move forward.
Goals
Help users make smarter decisions, faster
DealHunter83
2h ago
21 Posts
52 Reputation
Joined July 2005
I bought the X90L last month, and honestly, the picture quality is solid, especially in a dark room. But I can’t help feeling it’s a bit overpriced when you compare it to TCL’s mini-LED options. Those have better contrast and brightness for less money. Still, if you want a Sony, this is a safe pick.
8
2
We set out to surface the best of the community, without the noise. The goal was to design an AI-powered comment summarizer that made it easier to understand the general sentiment, pros, and cons of a deal, so users could make informed decisions quickly.
Direction to Launch
Designing within AI’s limits, without limiting the experience
1. Exploring the AI Landscape
I partnered with the PM and ML engineer to understand user pain points, assess the AI landscape, and align on how this feature could spark a new wave of assistive tools on Slickdeals.
2. Understanding AI output constraints
We worked closely with engineering to understand output limitations, like summary length, frequency of updates, and content accuracy. These shaped our approach and ruled out things like real-time comment syncing or user-generated voting.
3. Defining when a summary should appear
We established that summaries would only appear on threads with at least 10 meaningful comments. If a summary wasn’t available, the module stayed hidden to avoid misleading users.
4. Testing early output through Slack
Using Slack tests and internal feedback, we explored formatting and tone. Users wanted clarity, not fluff, so I pushed for a TLDR format with bullet points and a pros and cons list to support faster, more confident decisions
5. Designing for trust
To support our skeptical, older-skewing user base, I kept the tone soft and human. The design included clear labeling, a link to the full thread, and a collapse toggle, making it feel helpful, not invasive.
6. Setting the tone for a more delightful Slickdeals
I partnered with engineering to build one of the first microinteraction-rich features on the site. A dynamic loading state, typing animation, and smooth transitions added personality and excitement, without compromising clarity.
Delivery
A lightweight summary with a big impact
We shipped a lightweight AI summary module above the comment section on high-volume deal pages.
The final experience included:
This marked a turning point for the site, introducing both AI utility and a more modern, delightful design language.
Outcome
Early wins, and a glimpse of what’s possible
We soft-launched the feature to internal teams and select community members, focusing on deals with high comment volume.
Early results from testing and feedback included:
While we weren’t tracking hard metrics at full scale yet, early engagement showed strong signs that the summarizer was solving the right problem, and doing it in a way that felt intuitive and helpful.
Reflection
Lessons Learned
Working on this project was one of the first times I got to design for AI, and it pushed me to think not just about what we were building but how it would be received. It was exciting but also unfamiliar territory. We were introducing something new to a very human, community-driven space. The challenge wasn’t just designing for usability, but designing for trust, clarity, and tone. I had to balance what users needed, what the tech could realistically do, and how to introduce it in a way that felt aligned with the future of Slickdeals. It reminded me that designing for emerging tech isn’t just about being forward thinking. It’s about being thoughtful and intentional every step of the way.
Up Next
Thanks for Reading
Thanks for taking the time to read this case study, I hope it gave you a glimpse into how thoughtful design can make AI more useful, human, and trustworthy.
If you're curious to see more, feel free to check out my other case studies:
The Problem
When too many comments become too much of a good thing
Slickdeals is powered by its community, and popular deals often spark hundreds of comments filled with insights. But that volume came at a cost. Users struggled to find answers to key questions that impacted their decision making. Critical information got buried, and many gave up before getting what they needed. What was meant to build trust and connection had become overwhelming noise, causing frustration and drop-off instead of helping users move forward.
Goals
Help users make smarter decisions, faster
DealHunter83
2h ago
21 Posts
52 Reputation
Joined July 2005
I bought the X90L last month, and honestly, the picture quality is solid, especially in a dark room. But I can’t help feeling it’s a bit overpriced when you compare it to TCL’s mini-LED options. Those have better contrast and brightness for less money. Still, if you want a Sony, this is a safe pick.
8
2
We set out to surface the best of the community, without the noise. The goal was to design an AI-powered comment summarizer that made it easier to understand the general sentiment, pros, and cons of a deal, so users could make informed decisions quickly.
Direction to Launch
Designing within AI’s limits, without limiting the experience
1. Exploring the AI Landscape
I partnered with the PM and ML engineer to understand user pain points, assess the AI landscape, and align on how this feature could spark a new wave of assistive tools on Slickdeals.
2. Understanding AI output constraints
We worked closely with engineering to understand output limitations, like summary length, frequency of updates, and content accuracy. These shaped our approach and ruled out things like real-time comment syncing or user-generated voting.
3. Defining when a summary should appear
We established that summaries would only appear on threads with at least 10 meaningful comments. If a summary wasn’t available, the module stayed hidden to avoid misleading users.
4. Testing early output through Slack
Using Slack tests and internal feedback, we explored formatting and tone. Users wanted clarity, not fluff, so I pushed for a TLDR format with bullet points and a pros and cons list to support faster, more confident decisions
5. Designing for trust
To support our skeptical, older-skewing user base, I kept the tone soft and human. The design included clear labeling, a link to the full thread, and a collapse toggle, making it feel helpful, not invasive.
6. Setting the tone for a more delightful Slickdeals
I partnered with engineering to build one of the first microinteraction-rich features on the site. A dynamic loading state, typing animation, and smooth transitions added personality and excitement, without compromising clarity.
Delivery
A lightweight summary with a big impact
We shipped a lightweight AI summary module above the comment section on high-volume deal pages.
The final experience included:
This marked a turning point for the site, introducing both AI utility and a more modern, delightful design language.
Outcome
Early wins, and a glimpse of what’s possible
We soft-launched the feature to internal teams and select community members, focusing on deals with high comment volume.
Early results from testing and feedback included:
While we weren’t tracking hard metrics at full scale yet, early engagement showed strong signs that the summarizer was solving the right problem, and doing it in a way that felt intuitive and helpful.
Reflection
Lessons Learned
Working on this project was one of the first times I got to design for AI, and it pushed me to think not just about what we were building but how it would be received. It was exciting but also unfamiliar territory. We were introducing something new to a very human, community-driven space. The challenge wasn’t just designing for usability, but designing for trust, clarity, and tone. I had to balance what users needed, what the tech could realistically do, and how to introduce it in a way that felt aligned with the future of Slickdeals. It reminded me that designing for emerging tech isn’t just about being forward thinking. It’s about being thoughtful and intentional every step of the way.
Up Next
Thanks for Reading
Thanks for taking the time to read this case study, I hope it gave you a glimpse into how thoughtful design can make AI more useful, human, and trustworthy.
If you're curious to see more, feel free to check out my other case studies:
Work
Play
About
Resume
Tiffany Lim
Work
Play
About
Resume
The Problem
When too many comments become too much of a good thing
Slickdeals is powered by its community, and popular deals often spark hundreds of comments filled with insights. But that volume came at a cost. Users struggled to find answers to key questions that impacted their decision making. Critical information got buried, and many gave up before getting what they needed. What was meant to build trust and connection had become overwhelming noise, causing frustration and drop-off instead of helping users move forward.
Goals
Help users make smarter decisions, faster
DealHunter83
2h ago
21 Posts
52 Reputation
Joined July 2005
I bought the X90L last month, and honestly, the picture quality is solid, especially in a dark room. But I can’t help feeling it’s a bit overpriced when you compare it to TCL’s mini-LED options. Those have better contrast and brightness for less money. Still, if you want a Sony, this is a safe pick.
8
2
We set out to surface the best of the community, without the noise. The goal was to design an AI-powered comment summarizer that made it easier to understand the general sentiment, pros, and cons of a deal, so users could make informed decisions quickly.
Direction to Launch
Designing within AI’s limits, without limiting the experience
1. Exploring the AI Landscape
I partnered with the PM and ML engineer to understand user pain points, assess the AI landscape, and align on how this feature could spark a new wave of assistive tools on Slickdeals.
2. Understanding AI output constraints
We worked closely with engineering to understand output limitations, like summary length, frequency of updates, and content accuracy. These shaped our approach and ruled out things like real-time comment syncing or user-generated voting.
3. Defining when a summary should appear
We established that summaries would only appear on threads with at least 10 meaningful comments. If a summary wasn’t available, the module stayed hidden to avoid misleading users.
4. Testing early output through Slack
Using Slack tests and internal feedback, we explored formatting and tone. Users wanted clarity, not fluff, so I pushed for a TLDR format with bullet points and a pros and cons list to support faster, more confident decisions
5. Designing for trust
To support our skeptical, older-skewing user base, I kept the tone soft and human. The design included clear labeling, a link to the full thread, and a collapse toggle, making it feel helpful, not invasive.
6. Setting the tone for a more delightful Slickdeals
I partnered with engineering to build one of the first microinteraction-rich features on the site. A dynamic loading state, typing animation, and smooth transitions added personality and excitement, without compromising clarity.
Delivery
A lightweight summary with a big impact
We shipped a lightweight AI summary module above the comment section on high-volume deal pages.
The final experience included:
This marked a turning point for the site, introducing both AI utility and a more modern, delightful design language.
Outcome
Early wins, and a glimpse of what’s possible
We soft-launched the feature to internal teams and select community members, focusing on deals with high comment volume.
Early results from testing and feedback included:
While we weren’t tracking hard metrics at full scale yet, early engagement showed strong signs that the summarizer was solving the right problem, and doing it in a way that felt intuitive and helpful.
Reflection
Lessons Learned
Working on this project was one of the first times I got to design for AI, and it pushed me to think not just about what we were building but how it would be received. It was exciting but also unfamiliar territory. We were introducing something new to a very human, community-driven space. The challenge wasn’t just designing for usability, but designing for trust, clarity, and tone. I had to balance what users needed, what the tech could realistically do, and how to introduce it in a way that felt aligned with the future of Slickdeals. It reminded me that designing for emerging tech isn’t just about being forward thinking. It’s about being thoughtful and intentional every step of the way.
Up Next
Thanks for Reading
Thanks for taking the time to read this case study, I hope it gave you a glimpse into how thoughtful design can make AI more useful, human, and trustworthy.
If you're curious to see more, feel free to check out my other case studies: