OKX Assistant

OKX Assistant

Empower every individual to trade with confidence, learn smarter, rely on trusted support, and more

Empower every individual to trade with confidence, learn smarter, rely on trusted support, and more

OKX Assistant - Hero
OKX Assistant - Hero
OKX Assistant - Hero

Problem

Problem

One of the biggest challenges OKX faced was low customer support satisfaction. It’s an issue that directly impacts brand trust and reliability that’s critical in the fintech industry. This was a clearly defined problem that I was specifically hired to address. Over time, I led the team in evolving the solution into an all-in-one AI Assistant that not only improved support efficiency but also strengthened user trust and the overall platform experience.

Details

Details

My role: Lead designer

Duration: 3 quarters

Responsibilities: End-to-end process

Tools

Tools

  • Midjourney

CHAPTER 1

Discovery

research

Current experience

From my initial conversations with key stakeholders, I learned that the existing customer support experience was limited and faced challenges from both user and business perspectives. Users were frustrated by poor intent recognition and an overall subpar experience. From a business standpoint, the system suffered from limited component flexibility and poor integration with other in-house systems. Scalability was also a concern, and the solution was expensive to maintain.

CHAPTER 1

Discovery

Research

Current experience

From my initial conversations with key stakeholders, I learned that the existing customer support experience was limited and faced challenges from both user and business perspectives. Users were frustrated by poor intent recognition and an overall subpar experience. From a business standpoint, the system suffered from limited component flexibility and poor integration with other in-house systems. Scalability was also a concern, and the solution was expensive to maintain.

Highlights

Highlights

Highlights

The most significant user value came from the first phase of work, which focused on enhancing customer support

The most significant user value came from the first phase of work, which focused on enhancing customer support

Highlights
Highlights
Highlights

56%

56%

Increase in user satisfaction

64%

64%

Growth in self-service adoption

CHAPTER 1

Discovery

Research

Current experience

From my initial conversations with key stakeholders, I learned that the existing customer support experience was limited and faced challenges from both user and business perspectives. Users were frustrated by poor intent recognition and an overall subpar experience. From a business standpoint, the system suffered from limited component flexibility and poor integration with other in-house systems. Scalability was also a concern, and the solution was expensive to maintain.

CHAPTER 3

Design

CHAPTER 3

Design

CHAPTER 2

CHAPTER 2

Branding

Branding

CHAPTER 3

Design

CHAPTER 4

CHAPTER 4

Evaluative research

Evaluative research

RESEARCH

Usability test

Usability test

In the evaluative stage, I focused on testing one of the most frequently reported user issues: deposits not being credited. Through testing multiple variants, several key insights emerged:

  • Clear access to human support is critical in high-stake scenarios such as missing deposits. Delayed or unclear escalation increases anxiety and frustration

  • Early reassurance is essential in high-stress situations. Users need immediate confirmation that their funds are safe and an explanation that processing may take time

  • It’s important to avoid unnecessary requests (e.g., asking for the deposit method)

CHAPTER 5

Launch

CHAPTER 5

Launch

CHAPTER 6

CHAPTER 6

Internationalisation

Internationalisation

optimization

Experience at scale

Experience at scale

Throughout my design work, I supported internationalisation and localisation across Asia, Europe, and the Americas.

I ensured experiences were tailored to each market’s cultural nuances especially bridging the significant differences between our largest market, China, and the rest of the world.

CHAPTER 7

CHAPTER 7

Long-term vision

Long-term vision

WORKSHOP

Alignment sessions

Alignment sessions

After completing the ongoing initiatives, I transitioned to a more proactive role. I facilitated brainstorming sessions on how we could better serve users and establish a clear long-term direction. The sessions aligned the team on our vision, principles, and key ideas for the next two years.

CHAPTER 8

CHAPTER 8

Iterations

Iterations

OTC trading

OTC trading

OTC trading

Enable users to discover and trade tokens through a conversational experience

Enable users to discover and trade tokens through a conversational experience

All-in-one assistant

All-in-one assistant

Help users trade, learn, and get support at every step of their financial journey.

All-in-one assistant

Help users trade, learn, and get support at every step of their financial journey.

Outcomes

Outcomes

Outcomes

64%

64%

Growth in self-service adoption

35%

35%

Monthly active users

56%

56%

Increase in user satisfaction

High

High

Revenue Impact (Confidential)

Retrospective

Retrospective

What I would do differently

What I would do differently

One missed opportunity was spending more time shadowing customer service representatives early in the process to experience user problems firsthand. While I wouldn’t change the design decisions in hindsight, direct exposure to real conversations often surfaces nuances that don’t show up in dashboards or research summaries. Seeing these moments live can sharpen judgment, build empathy, and inform decisions beyond what data points alone can capture.

One missed opportunity was spending more time shadowing customer service representatives early in the process to experience user problems firsthand. While I wouldn’t change the design decisions in hindsight, direct exposure to real conversations often surfaces nuances that don’t show up in dashboards or research summaries. Seeing these moments live can sharpen judgment, build empathy, and inform decisions beyond what data points alone can capture.

Lessons

Lessons

A major lesson for me was adapting to new tools. The design landscape is changing rapidly, and on this project I found myself spending nearly 20% of my craft time coding and designing awith Claude rather than mapping flows and polishing details in Figma. This shift helped me iterate faster and explore more possibilities, but also required rethinking how I maintain design quality and consistency

A major lesson for me was adapting to new tools. The design landscape is changing rapidly, and on this project I found myself spending nearly 20% of my craft time coding and designing awith Claude rather than mapping flows and polishing details in Figma. This shift helped me iterate faster and explore more possibilities, but also required rethinking how I maintain design quality and consistency

Tradeoffs

Tradeoffs

The biggest trade-off was choosing between a fragmented experience and an all-in-one Assistant. It became a constant balance between execution speed and maintaining a clear long-term vision. While I knew the fragmented entry points weren’t ideal, I helped the team align on a longer-term direction and define the incremental steps needed to gradually move toward the ideal experience.

Get in touch

Feel free to reach out if you want to have a chat, or have any questions

Copyright © 2025 Konrad Marzec. All rights reserved.

Get in touch

Feel free to reach out if you want to have a chat, or have any questions

Copyright © 2025 Konrad Marzec. All rights reserved.

Get in touch

Feel free to reach out if you want to have a chat, or have any questions

Copyright © 2025 Konrad Marzec. All rights reserved.