LG MAL
Improving Label Accuracy and Brand Consistency with an AI-Powered Tool at LG
LG MAL
Improving Label Accuracy and
Brand Consistency with an AI-Powered Tool at LG
LG MAL
Improving Label Accuracy and Brand Consistency with an AI-Powered Tool at LG
LG MAL
Improving Label Accuracy and
Brand Consistency with an AI-Powered Tool at LG
01.
01.
intro
intro
My Role
Product Designer
TEAM
1 PM (HQ), 1 UXD (HQ), 4 Dev (HQ)
REGIONS
Korea, Seoul
US, New Jersey
US, Chicago
SKILLS
Product Design
Stakeholder Management
Interactive Prototyping
User Research & Testing
DURATION
Aug 2023- Jul 2024
In compliance with my confidentiality agreement, I have omitted and altered any confidential details.
Designed an internal tool to reduce errors and streamline workflows across LG’s global teams. As a core product designer, I improved UX writing quality, automated content reviews, and helped cut manual work between teams in Korea and the U.S.
Designed an internal tool to reduce errors and streamline workflows across LG’s global teams. As a core product designer, I improved UX writing quality, automated content reviews, and helped cut manual work between teams in Korea and the U.S.
In compliance with my confidentiality agreement, I have omitted and altered any confidential details.
SHIPPED
02.
02.
Problem
Problem
As LG positioned itself as a premium brand in the U.S., poor UX writing emerged as a major issue. Grammar mistakes, awkward phrasing, and poorly formatted text frustrated users and undermined the brand’s reputation for quality.
“Feels like a dumb AI translated it from another language into English.”
“Feels like a dumb AI translated it from another language into English.”
Google user review
Google user review
User Interviews
Interviews with global teams uncovered a key issue: manual checks slowed workflows and eroded confidence in the process.
Journey Mapping
Identified key emotional friction, and workflow inefficiencies across multiple touchpoints. It aligned cross-functional teams around user needs and informed design priorities.
Data Audit
Identified systemic issues: version control, grammar mistakes, inefficient feedback loops
03.
Research
DIVISION 3
DESIGNER
HQ KOREA
DIVISION 3
PM
HQ KOREA
DIVISION 2
DESIGNER
TECHNICAL WRITER
HQ KOREA
HQ US
DIVISION 2
PM
back and forth
though email
until satisfied
HQ KOREA
DIVISION 1
DESIGNER
HQ KOREA
DIVISION 1
PM
HQ KOREA
Final Approval
From Korean PM
[PRONE TO HUMAN ERROR]
[MAJORBOTTLE NECK]
[MAJORBOTTLE NECK]
untangling the problem
Interviews with global teams revealed that manual checks slowed workflows and hurt trust. Journey mapping uncovered emotional friction and inefficiencies across touchpoints, helping align teams around user needs. A data audit exposed version control issues, grammar mistakes, and broken feedback loops—highlighting a complex, fragmented English label process that spanned multiple teams and regions.
The English label process was more complex than expected—drafted in Korean at HQ, translated by different teams, reviewed by U.S. writers, then sent back to Korea for approval. The result was a fragmented, time-consuming workflow.
03.
Research
04.
Users
Understanding the distinct needs of each user group was key. The platform serves three primary users—Product Designers, Technical Writers, and Product Managers—each with unique goals, responsibilities, and workflows.
CREATE
KOREA HQ
HEEJAE
Product Designer
"We're not allowed to use external translation tools, so I constantly worry about grammar mistakes."
Fast verification of grammar
Easy search through past label examples
NEED
VERIFY
US HQ
MEGAN
Technical Writer
"I don't have time to version control my edits. I often end up sending contradictory changes, and it's really embarrassing."
NEED
MANAGE
KOREA HQ
JUNG HO
Product Manager
"Since different teams manage the label lists, the past labels aren’t stored in one place, making it hard for me to find the latest version for reference. "
Easy Search through guidelines
Access to past label examples
Universal storage for all labels and guidelines
NEED
Easy access to guidelines
Universal storage for edits
05.
04.
Goal
Goal
With tight timelines and growing stakeholder attention, the team quickly aligned on a goal: build an intuitive, web-based tool to centralize label workflows, enforce writing standards, and reduce back-and-forth between teams.
To achieve this, we distilled our focus into four key principles:
Centralize
label creation and review tools
SIMPLIFY
workflows across time zones
AUTOMATE
grammar/style checking with AI
REDUCE
human error and revision time
Centralize
label creation and review tools
SIMPLIFY
workflows across time zones
AUTOMATE
grammar/style checking with AI
REDUCE
human error and revision time
With tight timelines and growing stakeholder attention, the team quickly aligned on a goal: build an intuitive, web-based tool to centralize label workflows, enforce writing standards, and reduce back-and-forth between teams.
To achieve this, we distilled our focus into four key principles:
06.
05.
Solution
Solution
Rolled out MVP with three key features:
SEARCH
VERIFY
GUIDELINES
Rolled out MVP with three key features:
SEARCH
VERIFY
GUIDELINES
Search
Search to quickly find past label examples using a single word, sentence, or branded term, ensuring alignment with brand standards and accurate usage.
Search to quickly find past label examples using a single word, sentence, or branded term, ensuring alignment with brand standards and accurate usage.
Verification
Verify feature allow users to review entire labels, leveraging Busan University AI to catch minor grammar errors and ensure compliance with universal style guidelines.
Verify feature allow users to review entire labels, leveraging Busan University AI to catch minor grammar errors and ensure compliance with universal style guidelines.
Guidelines
Allowing user to access the latest UX writing guidelines across all HQ divisions in one platform, eliminating the need to track down document owners for product term details.
The guideline library feature allows users to access the latest UX writing guide across all HQ divisions in one platform, eliminating the need to track down document owners for product term details.
07.
User Test
The platform launched as a three-month pilot, used daily by over 100 HQ team members who provided ongoing feedback. We also tested the English interface with five users in NJ and five in Chicago to ensure clarity. Developers implemented feedback in real time, allowing for fast, continuous improvements.
06.
User Test
The platform launched as a three-month pilot, with over 100 HQ content team members using it daily and providing ongoing feedback. We also tested the MVP with five internal users in NJ and five in Chicago, focusing on the English interface to ensure clarity and intuitive UX writing. Developers applied feedback in real time, enabling rapid iteration and continuous improvement.
08.
Final Iteration
Mass
Verification
Korea HQ users requested a bulk verification feature to boost efficiency. We added Excel upload and HTML page verification to simplify tasks and improve workflow.
Korea HQ users requested a bulk verification feature to boost efficiency. We added Excel upload and HTML page verification to simplify tasks and improve workflow.
07.
Final Iteration
UX Writing
Unmoderated testing showed users were confused by the UX writing. We fixed this by simplifying labels, terms, and language to make the platform more intuitive.
Unmoderated testing showed users were confused by the UX writing. We fixed this by simplifying labels, terms, and language to make the platform more intuitive.
09.
08.
Impact
Impact
134 Error Detected
The system identified 134 errors in published app versions—nearly three times more than the 51 caught through manual review.
380 Errors Pre-ship
In 2022, 37 errors were caught before shipment; during the 2023 pilot, the same product surfaced 380 errors.
50% Time Saved
The tool cut revision time by 50%, allowing teams to deliver polished content faster and more efficiently.
134 Error Detected
The system identified 134 errors in published app versions—nearly three times more than the 51 caught through manual review.
380 Errors Pre-ship
In 2022, 37 errors were caught before shipment; during the 2023 pilot, the same product surfaced 380 errors—highlighting the effectiveness of the automated system.
50% Time Saved
The tool cut revision time by 50%, allowing teams to deliver polished content faster and more efficiently.
10.
Reflection
This project underscored the value of clear communication and fast, flexible execution in delivering meaningful solutions. Designing a universal content system for LG required constant prioritization—balancing evolving global needs with tight deadlines and limited resources.
What stood out most was how small, thoughtful changes added up—each iteration brought us closer to solving a much larger problem.
The work was especially rewarding when the project was recognized with LG’s Internal Innovation Award for its cross-functional impact.
09.
Reflection
This project underscored the value of clear communication and fast, flexible execution in delivering meaningful solutions. Designing a universal content system for LG required constant prioritization—balancing evolving global needs with tight deadlines and limited resources.
What stood out most was how small, thoughtful changes added up—each iteration brought us closer to solving a much larger problem.
The work was especially rewarding when the project was recognized with LG’s Internal Innovation Award for its cross-functional impact.
Enter Password
OTHER PROJECTS
OTHER PROJECTS
