Dcard AI Lab
Where Ideas Transcend, Solutions Emerge

ai.dcard.io
About the project
Summary
The AI Lab at Dcard serves as an internal hub for AI tools, use Chat GPT technology to create a brand new interface and platform. This initiative involves accumulating high-quality prompts for the company. and seamlessly integrating AI capabilities into our workflow. This empowers our colleagues to optimize their work processes, enhancing both efficiency and quality.
In this project, I took on a dual role as a PM and Designer, collaborating with the team to bring it to fruition within a month. Once launched, it significantly facilitated access to AI tools for our colleagues, using in various aspects of our work, such as translation tasks, drafting multilingual emails, generating copy and content ideas, and reviewing regulations. The project has successfully achieved the accumulation of prompts, promoting their reuse, and enhancing work quality and efficiency. AI technology has now become a reliable partner in our day-to-day operations.
My Role
Product Manager & Product Designer
Product Management, User research, Interaction, UI Design, UX Writing, Prototyping
Plat Form
Web
Collaborative Partner
Engineer, Marketing & Operation departments
Project Time
2023.05-2023.06 ( 1 month)
Website
Project Hypothesis & Goals
Dcard is a highly penetrated and influential social platform among the young generation. Apart from Taiwan, it operates in Hong Kong, Japan, and Peru. Given the forum-based nature of our product, there's a need for extensive content generation, copywriting, and multi-language communications in our operations. While leveraging AI technology, we are developing user-friendly AI tools to streamline our work content and processes. We are also storing AI prompts, which will be instrumental in assisting us expand into international markets and reduce the cold start-up time when entering new markets in the future.

The Progress
The goal of this project is to complete and launch a brand new internal AI tool within one month. I took on the roles of both Project Manager (PM) and Product Designer (PD), which involved tasks such as defining product workflows and design, defining feature logic, designing interactions, conducting user interviews and testing, and more.
In the initial planning phase of this project, we conceptualized three different versions: Basic, Intermediate, and Ideal. We then collaboratively identified executable tasks based on scheduling requirements and team discussions. Additionally, our engineering team research on AI integration, design UIUX in the same time. By discussing and do the user tests frequently, the project launched within one-month successfully.

Design
After outlining the project goals and phases, we proceeded to create multiple versions of the editing interface using wireframes. We then swiftly developed prototypes for internal user testing and interviews to identify the most suitable version. During these interviews, we gathered feedback indicating the need for the ability to view original prompts at any time, enabling quick access for review and modification in case the AI's performance was subpar. We also included a feature to record response iterations, allowing users to track answers after multiple interactions. Furthermore, we enhanced the copy button for easy copying and pasting, and incorporated a GPT version toggle, giving our colleagues the flexibility to choose based on their specific requirements.​

Three design strategy of AI page

List of AI tool
Edit Page

AI Operation Page
Page - See the Prompt

ai.dcard.io
Takeaways & Learning
1. Rapid Decision-Making Under Multiple Constraints:
-
This project operates under a tight schedule, requiring the construction of an internal AI tool from scratch within a short timeframe. The challenge lies in quickly aligning the thoughts and goals of company stakeholders and defining the basic, intermediate, and ideal feature versions through discussions with the team.
2. Small-Scale Design Validation:
-
Due to time constraints, we adopted existing development frameworks and packages to rapidly create wireframes or initial designs. Multiple versions were produced, and simple user interviews were conducted internally to swiftly converge on suitable design layouts and operational modes.
​
3. Subsequent Tracking and Observation:
-
The initial version of the project serves as the basic version, primarily aimed at familiarizing company colleagues with the AI tool. Continuous monitoring of usage patterns, regular feedback collection, and planned extensions of advanced features ensure the ability to make rapid optimizations in the future (e.g., command categorization, favorites, adjustments for neutral or extreme responses, etc.).
4. Time Management with Dual Roles as PM & PD:
-
The project's dual roles as PM and PD highlight the critical importance of time management. Advanced planning outlines different stages of product outcomes and, after initial discussions, converges on high-priority features to ensure a smooth on-time launch within the schedule.
*The intellectual property rights of the content in this project are owned by Dcard.