Coach AI background

CAREERVILLAGE'S COACH AI — MOBILE

Redesigning Onboarding for AI-Driven Career Guidance

Coach, CareerVillage's AI career advising chatbot, suffers from a disjointed onboarding process split into sign-in, survey, and chatbot sections. This redesign transformed Coach into a cohesive, engaging presence, boosting user satisfaction by 22% as more users completed the onboarding process.

My Role

UX Designer

The Team

Xingchen Cao / Cody Mak / Wendi Zheng

Timeline

5 weeks

Tools

Figma / Google Drive / Slack

Client ProjectMobile Onboarding

3.5% Conversion Rate: Overhauling Coach AI's Disjointed Onboarding Process

In 2024, Coach launched an AI chatbot designed to deliver personalized career guidance based on user preferences from onboarding. However, the current onboarding process was underperforming, with a dismal 3.5% conversion rate.

Coach AI chatbot before
Coach AI survey before

Revitalizing Coach's Onboarding for a Fun, Engaging Experience

The goal is to revitalize Coach's onboarding experience, making it more fun, engaging, and user-friendly. This includes streamlining and unifying the three disjointed sections into a seamless, cohesive flow. Our constraints included WCAG 2.0 AA color compliance and maintaining the original survey question wording.

Increase conversion rate of 3.5%Increase user satisfactionIncrease user completion

HMW Question

“How might we make Coach AI fun and engaging to cater towards career professionals, pivoters, and students to support their professional development?”

Uncovering Pain Points: Insights from Comparative Analysis & User Interviews

We conducted a comparative analysis of onboarding processes from leading platforms like Duolingo, Stoic, Grammarly, Wysa, and Handshake.

Duolingo onboarding

Duolingo

Inspired the use of a character/persona

Wysa onboarding

Wysa

Concise, highly engaging chatbot messages

Stoic onboarding

Stoic

Personalized loading page

User Interview Insights

User 1

The survey sliders were challenging to use and frustrating to interact with.

User 2

The onboarding process felt unnecessarily lengthy and drawn out.

User 3

I'm curious about what sets Coach AI apart from other generative AI platforms like ChatGPT.

Restructuring Onboarding: Survey First, Chatbot Last for Better Engagement

We kept the overall user flow the same but shifted the chatbot to the end of the onboarding process. Based on user feedback, we decided to prioritize the survey at the start to better engage users, while still giving them a preview of the chatbot's functionality later on.

Coach AI process flow

Streamlining Onboarding: Rapid Wireframing for a Fresher Approach

For the lo-fi wireframes, we began by sectioning off the onboarding process flow. Given the tight timeline, we jumped straight into wireframing and brainstorming new approaches.

Character Iteration

Character iteration wireframes

Survey Scale Iteration

The original survey scale was challenging to interact with, so we created several iterations to make it more fun and compelling.

Survey scale iterations

Chatbot Iteration

We shortened the chatbot to align with the tutorial, allowing users to familiarize themselves with the tool before completing the onboarding.

Chatbot iterations

Coach AI Transformation: Characters, Emojis, Personalization, and Simplified Chatbot Flow

We completely transformed Coach AI's onboarding process by introducing characters to personalize the experience.

Coach AI top 3 designs
Before vs. After Survey Flow
Before vs. After Chatbot Flow

Ensuring Universal Appeal: Testing Coach AI's Character and Survey with Adult Users

We interviewed adults (ages 20–40) in the workforce or pursuing Master's/PhD programs. Key insights: users appreciated the character for giving the AI a distinct personality, but were unclear about the purpose of the survey.

User 1

I enjoyed the survey, but I'm curious how Coach uses my answers to personalize the onboarding process.

User 2

I really liked the emoji scales in the survey because they made me want to answer the questions.

User 3

I liked the bear; it made the AI feel personal and gave it a real personality, not just 'nothingness technology.'

22%

increase in user satisfaction

User Satisfaction Increased by 22%: Client to Launch Redesigned Onboarding in Q2 2025

Based on participant interviews and testing, we increased user satisfaction rate by 22%, as users expressed more willingness to complete the onboarding process with our redesign. The client plans to ship the product in Q2 2025 after reviewing the prototype with her team and executive leadership.

Challenges

Working with our client was a great experience, though we faced challenges balancing scope and timeline to create a fun, engaging onboarding design. Incorporating a persona without overshadowing the survey was tricky, and designing a measurable survey scale required careful adjustments.

If We Had More Time

We would add more actions and emotional states to the character, include a message before the survey to clarify its purpose and benefits, and conduct further user testing for consistency.

Team photo