Developing
Beyz 2.0 Planning
AI Interview Copilot
HomePage
Create New Copilot
Copilots Dashboard
Settings
...
AI Interview Copilot
HomePage
Create New Copilot
Copilots Dashboard
Settings
...
Other Features
Core Features
Higher priority
Core Features Beta Test
Core
Feature Design
Core Features Refine
Other Features Design
2.0 Launched
02/29/2024
01/29/2024
All
Features Beta Test
How we got here and where we are now
Iteration Path
MVP
R&D
Design
Test
Development
Launch
Beyz 1.0
Beyz 1.1
Beyz 2.0
Sep 2023
Mar 2023
Why fear interviews when you can practice with AI? The AI Interview Copilot offers real-time, AI-driven answer suggestions to help you tackle any question with confidence. Get ready to impress with every answer.
Introduction
Focus Area
UI&UX Design
Product Iteration
User Growth
Tools
Figma
Photoshop
OpenAI API
Role
Co-founder
UI&UX Designer
Product Manager
Timeline
Sep 2023 - Present
How did I get here?
Launched on Beyz.One
Beyz
An AI driven career enhancement tool with over 3,000 registered users
Explore Our Product
How we got here and where we are now
Highlighted Features
Instant AI Customization (Live demo)
Quick Create streamlines the process of getting started with our platform, allowing users to swiftly input their details and jump straight into preparing for their interviews. This feature reduces the initial setup time significantly, enabling users to focus more on what truly matters—honing their interview skills.
Introduction
AI Interview Copilot (Live demo)
The AI Interview Copilot is your personal guide through the daunting world of job interviews. Leveraging cutting-edge AI, this feature provides real-time, tailored advice and suggestions, helping you navigate through each question with confidence. It's like having an expert interview coach by your side, anytime and anywhere.
Product Iteration
Streamlining
Success
We launched with a basic MVP, quickly iterating to versions 1.0 and 1.1 based on direct user feedback. Through these rapid iterations, we've continually optimized our product to better meet our users' needs.
MVP
UI Iteration
Real-World Runs (Vercel)
In our MVP version, users pointed out the lack of consideration for previously uploaded resumes. There's a clear desire for a more efficient way to reuse information already entered, suggesting improvements towards streamlining the resume upload process to better accommodate returning users.
V 1.0
UI Iteration
Real-World Runs
Feedback on Version 1.0 highlighted that the answer analysis reports were overly lengthy and dense, deterring users from reading through each section. Users expressed a need for concise, well-organized content and requested summaries of how their responses could align more closely with ideal answers.
V 1.1
UI Iteration
Real-World Runs
In Version 1.1, users found the 'View Hint' feature's sliced answers to be too fragmented and lacking in meaningful guidance. There's a strong call for reformatting hints into more comprehensive, order-specific phrases that offer clear, actionable insight into crafting responses.
UI Iteration Sample
UI Iteration
04/11/2023
MVP
03/01/2024
V 1.1
14/12/2023
V 1.0
Stage review takeaways
What does our product face?
Limited Product Timeliness
Insufficient User Trust
Weak Demand Rigidity
Most users start prepping only after receiving an interview invitation, which usually leaves them with just a few days to a few weeks to prepare. As a result, our product's applicability is narrow, with limited timeliness.
Given the high stakes of interviews, users are often skeptical about adopting new tech-driven tools, hesitant to invest their precious prep time in unfamiliar solutions.
Even though traditional solutions typically lack sufficient personalization, they still enable users to complete their interview prep, diminishing the irreplaceability of new products.
Addressing User Inertia
Facing the reality of user inertia head-on, the reliance on 'old ways' to slowly improve oneself offers limited intrinsic motivation for most users.
Leveraging the Importance Mindset
Capitalizing on users' serious attitude toward interviews, we provide a killer product that can effortlessly boost their interview performance in a short period.
Creating a Competitive Edge
By fully leveraging the advantages of AI, we differentiate our product in the market, breaking through users' 'tech skepticism' with compelling product stickiness.
advantage?
How should we turn it to our
challenges
Beyz 1.1 · Live Demo
See how we revamped our product to win over our first loyal users in Phase 2
Quick Building
Highlight Features
Customized
Questions
Tailored specifically to your resume and job application, our AI crafts targeted interview questions to prepare you effectively.
Highlight Features
Answer
Guidance
Offering suggested responses and key point prompts for interview questions, helping you formulate more polished answers.
Highlight Features
Practice and
Feedback Reports
Enables practice through voice and text responses. Submitted answers are analyzed to generate feedback reports, highlighting strengths and areas for improvement.
Context
MVP Design
Blueprinting Innovation
After gathering feedback, I dove into a round of ideation, sketching out the first version of a low-fidelity MVP UI. This blueprint, brought to life through Figma prototypes, was my first tangible step towards realizing our vision.
MVP User Test & Modify
Refining Through
Feedback
With my MVP UI in place, I embarked on multiple rounds of user testing via Vercel (Native Prototyping). Each cycle brought invaluable insights, allowing me to rapidly iterate and optimize, ensuring the product not only met but exceeded user expectations.
Ideation & Low-Fi Prototype
User Interview doc
Stakeholder Meetings
Aligning
Visions
I kicked off our journey with several stakeholder meetings, aligning on project goals and expectations with key team members and investors, setting the stage for targeted development.
Stakeholder Meetings
College students, especially those nearing graduation, often lack sufficient preparation and understanding for interviews. Many possess relevant professional experience and backgrounds but feel puzzled about how to highlight their expertise and skills or how to answer specific questions.
Phase One
MVP Launched!
Phase Two
Context
After gathering and analyzing feedback from Phase 1, we embarked on Phase 2 to elevate our product. This phase introduces the "AI Copilot," integrating "AI Interview Assistant" and "Cheat Sheet" functionalities, aiming to revolutionize interview prep with deeper personalization and efficiency.
Brainstorming
Why...?
"Because they feel that compared to the time-consuming practice process, the training results don't directly contribute to their chances of interview success."
Why are users reluctant to spend a lot of time preparing for interviews with our product?
FIRST WHY
"Because they believe even with preparation, there's no guarantee they'll encounter the same questions during the actual interview."
Why do users think training results don’t directly help with success rates?
SECOND WHY
"Because improving articulation skills in a short time is challenging, and in an interview setting, facing different questions can easily lead to nervousness and difficulty in figuring out how to phrase responses."
Why do users think not encountering practice questions during the interview equals no help?
THIRD WHY
How might we leverage the capabilities of large language models in text comprehension and expression reconstruction to help users bypass the traditional, inefficient practice process?
How might we directly assist users in their thought process for answering questions in a real interview environment?
HMW:
HMW:
2.0 Core Features Design
Designed for
User Inertia
Design a tool that offers real-time answer suggestions in a true interview environment, helping them quickly master various question types and directly enhance their interview performance.
Prototyping
Part 01
Interview AI Assistant
Part 02
Hint Sheet
UI Modify
2.0 Core Features Usability Test
FeedBack
Summary
Usability tests reveal most users are unsure of how to initiate question listening and find full answers too extensive for quick review, hindering their effectiveness as prompts.
Before
After
Before
After
Incorporate more visually distinct icons in the startup background to hint at interactive features, and revamp the prompt generation rules by using bullet points instead of full answers to enhance readability.
Solution
Designed for
User Inertia
Design a tool that offers real-time answer suggestions in a true interview environment, helping them quickly master various question types and directly enhance their interview performance.
2.0 ‘Other Features’ Design
Prototyping
UI Sample
Design guidelines
2.0 ‘Other Features’ Usability Test
How do we solve this?
Vs
Enhanced product performance
Easier user onboardinge
To achieve better output, our product requires users to input detailed personal information in preliminary steps, yet this increases friction in the 'create new' process, leading to a higher dropout rate.
Prototyping
Friction
Reduction
2.0 Product UX modify
Entering the copilot and clicking "create new."
Users were led to an information input screen, required to fill in all mandatory fields for comprehensive AI personalization.
High Friction
Original Flow
Completion of creation.
Current Flow
Low Friction
Users are presented with a few quick multiple-choice questions.
Entering the copilot and clicking "create new."
Completion of creation.
I identified that the greatest friction was in step two. The vast amount of information required upon first entry significantly increased users' uncertainty and decreased their desire to explore further, notably increasing the dropout rate at this stage.
Analysis
Consequently, I distilled the essential information into three main categories—"Interview Type," "Answer Tone," and "Personal Introduction"—simplifying them into two multiple-choice questions and two short-answer questions, displaying only one item at a time. Now, a new copilot can be created in five steps or less, allowing users to quickly input information and begin exploring. If further personalization is desired, users can easily return to the dashboard to upload more personal details.
Reduced dropout rate by 30%
*The amount of information that needs to be filled out
Originally, users faced a cumbersome process filled with mandatory fields for AI personalization. Recognizing the friction, I simplified entry into just a few essential selections, drastically reducing setup time and encouraging further exploration.
Keys Takeaways
Reduce User Thinking
Aim to lessen the cognitive burden on users. Avoid requiring them to manage information in any form, even if it's a one-time task.
Test Whenever Possible
Development is costly. Wherever feasible, conduct tests to ensure that decisions are informed and resources are utilized effectively.
Users Dislike Losing Existing Content
Even if their previous experiences were suboptimal, users hate seeing existing feature removed. Maintaining familiarity and continuity is key to user satisfaction.
Continuously updated...
Discover how I led a project from concept to launch and achieved commercial success.