2024
Yukin.ai
Empowering Non-Tech Users: Knowledge Base Redesign
About Yukin
Yukin is an AI automation platform that offers chatbots, voice bots, and marketing bots to help businesses streamline customer support, boost sales, and run personalized campaigns.
About the Project
As Yukin grew, its chatbot platform began serving a diverse range of users—both technical and non-technical.
However, non-developers struggled to train and manage bots independently, often relying on engineers for basic tasks.
We aimed to redesign the platform with a self-serve approach, making bot training and customization intuitive for everyone.
Problem Statement
Non-technical users lacked the tools to train or personalize their bots without developer support. This created dependency, slowed down workflows, and limited the adoption of bots across teams. The absence of a user-friendly system made it difficult to manage bot knowledge and brand alignment at scale.
Reliance on Developers
Non-technical users had to depend on developers for bot training and customization, slowing down workflows and reducing autonomy.
Lack of Knowledge Management
There was no dedicated space to upload or organize knowledge, making it hard to train bots with accurate, updated information.
No Easy Customization
Users couldn’t visually personalize bots—like changing color, name, or behavior—without technical help, limiting brand alignment.
Key Performance Indicators
Before jumping into design solutions, we defined key metrics that would help us measure success.
Autonomy
Are non-technical users able to train and customize chatbots without developer assistance?
Adoption
How many users are actively using the Knowledge Base and Bot Customization features after launch?
Research
As Yukin’s customer base expanded, we noticed a recurring pattern in user feedback, particularly from non-technical users.
Support teams frequently received emails requesting help with tasks that ideally should be self-service. To better understand user needs, we analyzed support tickets, conducted lightweight interviews, and mapped out core friction points.
32%
of bot-related support tickets came from users asking how to train or customize their chatbot without developer help.
1 in 3 emails
we received was a request like “Can I change the bot’s color?” or “How do I update answers myself?” — highlighting a need for visual control and knowledge autonomy.
How we solve this?
How this works?
The Customize Agent panel allows users to easily modify each bot’s name, icon, color, and connect it to a specific Knowledge Base—all in one place.
Customize an agent
We designed the Agent Details view to offer an intuitive editing experience for bot administrators. Everything from branding to behavior is configurable—no coding required.
Knowledge Base
The transaction tab feature in our expense tracking app provides users with a comprehensive list of all their past transactions.
Problem with this Screen
After conducting five user testing sessions, we discovered that users struggled to identify knowledge bases by their unique IDs alone. To address this, we introduced a headline-based naming system to make them easier to recognize and manage.
Overall Impact
The spending feature of our expense tracking app consists of three tabs, providing users with a comprehensive overview of their finances.
OVERALL METRICS
+18%
User Retention
+26%
User Retention
+21%
User Retention
FEATURE DISCOVERY
+72%
Knowledge Base
+57%
Ticket resolving
+41%
Customization Usage