pixsee cover photo
overview

Designing AI Slideshow album to enhance user engagement with smart camera

pixsee is a smart baby monitor from Compal Electronics, which provides an auto-capture feature. However, what challenges do parents face when documenting their babies' lives? In this challenging but fascinating project, I conducted UX research and usability testing to improve the auto-capture feature of pixsee, and designed product features correspondingly.
holder
compal logo
timeline
Aug 2021 - Jan 2022
(6 months)
role
UX Designer
UX Researcher
team
UX Researcher / 3
UI Designer / 3
Product manager / 1
Software engineer / 3
AI engineer / 2
Sales & Marketing / 2
challenge & Goal
How can we enhance user engagement with the AI auto-capture feature of our baby camera?
The AI auto-capture feature is the primary selling point of our product and thus has high business value. However, our Google Analytics data reveals that user engagement with this feature has been lacking. The project seeks to uncover insights into:

1. What are users unsatisfied about with the auto-captured feature?
2. How can we enhance user engagement and foster loyalty towards our core automated photography feature?
data
solution overview

Auto-generated slideshow with face recognition capability

Created an auto-generated slideshow that gathers a series of existing photos for users, in which users can find and organize memories of their loved ones.
pre-made album
Get auto-generated slideshow videos regularly while retaining the ability to customize and edit them as needed.
full screen
face filter
search with face filters
Enable users to choose photos featuring the selected faces.
Maintain editing capabilities
Users maintain editing and customization capability for their albums, including adding music and reordering.
add music
design process
design process
01/ user research
I conducted diary studies, competitor analysis, and a survey to generate product strategies for developing new features and enhancing algorithms.
research goal
What are the needs that have not been met in terms of taking photos of their babies using AI camera?
  • How satisfied are users with the photos we push to them? If they express dissatisfaction, what specific content or precious moments do they prefer instead?
  • While recording precious moments using photos, what pain points have not been addressed? Can we add on to the current auto-capture features to provide additional value to users?
Research insights
1. The photos users receive are too similar and repetitive, which leads to low levels of interest.
Parents typically don't relocate the baby monitor once it's set up, resulting in consistently similar photos and subsequently, a lack of interest in viewing the album.
2. Users find it hard to select which moments captured they preferred
The types of photos that users like are highly diverse, and it is difficult to find a common ground to create a scoring rubric that meets everyone's expectations for AI.
3. Users are overwhelmed by the sheer number of photos sent to them by the baby camera, resulting in not checking them out.
Subscribing to the service entitles users to receive 12 photos daily. However, as users typically don't organize or clean up their photos every day, the accumulation of images in their albums results in information overload, ultimately leading to infrequent checks by users.

The challenge lies in striking a balance between ensuring users feel the service is worth their payment while also avoiding overwhelming them.
02/ Persona
primary persona
03/ Ideation
Auto-generated
slideshow
Gathers a series of existing photos for users to help them organize and rediscover past precious moments.

auto-gen slideshow success matrix
voice command
to take photos
Implementing voice-triggered capture lets users define their precious moments and increases the accessibility of the camera to retain the moment.
voice command success matrix
add auto-recording
to the two-way talk
Leverage current features and auto-records when parents use the 2-way talk that helps parents talk to their babies remotely to see their kids’ cute reactions.
auto-recording success matrix
Based on the success metrics, implementing auto-generated slideshow is prioritized since it
  • Meets users' needs;
  • Is easy to implement;
  • Can be a great marketing materials since it would be a brand new feature.
04/ user flow
user flow
05/ Mockups & iteration
Adjusted
Before
design alternative _ cover
Enlarge slideshow cover, eliminate create button
Emphasize the AI capacity by removing the create button and enlarging cover photos for marketing purpose.
design alternative_ play button
add a play button to better preview the edition
In the reorder section, a play button enables instant preview, ensuring order-music compatibility.
design alternative _ layout
Choosing a layout resembling Album
This decision aims to enhance consistency and increase user efficiency when searching for photos based on specific dates.
1. How can users access the slideshow feature?
option1:
Menu
side bar access
option2:
tab
tab access
Option3:
family album page
album display access
Following the usability heuristic that prioritizes recognition over recall, the access to features should not buried in layers, eliminating the need for users to remember intricate access pathways. Furthermore, access to the feature should be prominently evident to better serve the marketing purpose.
2. How many editing functionalities should we authorize for users?
Highlighting our AI capabilities and taking implementation time into account, users can currently utilize fundamental editing features like reordering and adding/deleting photos. Advanced functionalities such as adding texts/stickers, customizing photo lengths, and selecting specific music periods are slated for future implementation.
06/ prototypes
high fi prototype
impacts
  • Launched a new product feature through end-to-end design, bringing a win-win situation for both users and our business.
  • Collaborated with product managers, software engineers, the sales and marketing team, and third-party service providers to implement features.
lessons learned