Crypto Wallets
Ubiquitous Computing
Feature ideation
Team
Kanika Bansal
Deepak Mangapuram
Kavya Maragoni
Sneha Sivasubramanian
Timeline
Context
10 weeks
New and moderately experienced users of crypto wallets face challenges like complex terminology, intimidating onboarding processes, and opaque security practices which create cognitive overload and hinder wider adoption. The challenge lies in designing a wallet experience that simplifies these complexities while empowering users to navigate cryptocurrency confidently and securely.
Problem Statement
How might we enhance the trust, usability and security of Crypto Wallets, to increase the adoption rate?
Reduce User effort through invisibility
Seamless Interaction
Project Goals
A little background
“The severity of these threats is evident in the fact that scammers collected $679 million in cryptocurrency through various fraudulent activities in just the first half of 2024”
Other ongoing research efforts
Anomaly detection algorithms to identify suspicious transactions
Generative adversarial networks (GANs) to simulate and prepare for potential security breaches
Neural networks for analyzing complex fraud patterns
But
Many wallets lack effective user education tools, leaving users vulnerable to day to day usage and phishing scams. Additionally, the complex terminology in most crypto wallets create significant barriers for novice users, thereby effecting the adoption.
This is where CryptoBuddy steps in! CryptoBuddy is an AI-powered assistant designed to enhance the user experience of cryptocurrency wallets, particularly for novice users by:
Providing personalized, step-by-step guidance during the onboarding process
Offering context-aware assistance for complex tasks like wallet setup and fund transfers
Delivering real-time, adaptive support to users as they navigate the crypto wallet
Impact on user experience through CryptoBuddy
1.
Context- Awareness
Outcome: CryptoBuddy effectively is aimed to provide personalized, context-sensitive guidance during onboarding and transactions. It adapts its support based on user behavior, such as offering corrective feedback after repeated errors.
Impact: Users felt supported without being overwhelmed, as the AI assistant anticipated their needs and offered timely help, reducing confusion and building confidence.
2.
Calm Technology
Outcome: Notifications and guidance were designed to be non-intrusive, aligning with the principle of "calm technology." The AI assistant used subtle cues like tooltips or micro-interactions to provide information.
Impact: Users appreciated the balance between receiving helpful insights and maintaining control over their interactions, fostering trust in the system.
3.
Invisibility
Outcome: The AI assistant operated unobtrusively in the background, activating only when users encountered challenges or made mistakes.
Impact: This seamless integration minimized interruptions, allowing users to focus on their tasks while feeling reassured that assistance was readily available when needed.
4.
Outcome: Through clear explanations of crypto concepts and proactive error prevention, CryptoBuddy empowered users to manage their wallets confidently.
Impact: Users reported feeling more capable of navigating cryptocurrency tasks independently, which is critical for fostering long-term adoption.
Empowerment
Enough with the overview, let’s look at the journey! Shall we?
Secondary Research
User Interviews
User Flow Diagram
Storyboards
Our secondary research highlighted critical insights into the challenges and barriers associated with cryptocurrency adoption. As shown in the scenario summary:
We conducted interviews with 12 participants, including novice and intermediate cryptocurrency users. Questions focused on their experiences with cryptocurrency wallets, security concerns, and preferences for assistance during transactions.
In the process of developing the AI-powered CryptoSentinel wallet, we have explored various user scenarios to better understand and address the challenges that users might face in the cryptocurrency environment. To guide our ideation and design process, we've formulated several "How Might We" questions based on our desk research. These questions aim to focus our efforts on creating meaningful solutions that enhance user experience and security. Below are the How Might We questions derived from our storyboard scenarios:
Scenario 1
How Might We Ensure User Security When Connected to Unsecured Public Networks?
Scenario 2
How Might We Assist Newcomers in Understanding and Navigating the Cryptocurrency Space?
Scenario 3
How Might We Prevent Users from Falling Prey to Phishing Attacks and other scams?
Context and Research Background
Research Goals
Reduced User Effort Through Invisibility:
Leveraging AI to automate security tasks and integrate seamlessly into users’ routines.
Aligns with the ubicomp principle of invisibility, where the system operates in the background without active user intervention.
Seamless User Interaction:
Context-aware design to provide tailored assistance based on user behavior and environment.
Ensures security and usability in dynamic settings, such as connecting to public networks or managing complex crypto features.
Background and Context
Only 6.8% of the global population actively uses cryptocurrency
Users fear unauthorized access, scams, and the potential loss of digital assets, especially when using hot wallets.
Most users are unfamiliar with crypto which is one of the main cause of lower adoption
(Source: Statista Cryptocurrency Report, 2024).
(Source: Global Blockchain User Trends Report, 2023
(Source: Crypto Security Survey by Kaspersky, 2023).
Beginners particularly struggled to recognize warning signs, making them feel vulnerable.
Overwhelming Onboarding Process
Cognitive overload
Too many Jargons
Fear of Security Risks
Participants found the initial setup intimidating due to unfamiliar terminology and a lack of clear guidance.
Many abandoned the process mid-way or avoided crypto altogether.
Too many jargons were overwhelming
Users expressed concerns about scams, phishing attacks, and transaction errors.
Findings & Insights
Personas and Their Influence on Design
To ground our design decisions in real-world user needs, we developed two personas based on insights gathered from user interviews:
Conceptual Framework Overview
The design process for CryptoBuddy was informed by key ubiquitous computing principles: context-awareness, invisibility, and adaptability to create a cryptocurrency wallet experience that is intuitive, supportive, and minimally intrusive. These principles helped shape features that align seamlessly with the user’s needs, empowering them to manage cryptocurrency confidently while reducing cognitive load.
Our UX conceptual framework centered on two main goals:
Empowerment
Ease of use
Michael’s persona directly informed the creation of the Context-Aware AI Onboarding Assistant
This feature ensures step-by-step support during wallet setup, using plain language and non-intrusive prompts to provide confidence without overwhelming him.
The assistant was also designed to fade into the background as Michael became more comfortable, aligning with his preference for minimal intrusion
Jessica’s persona guided the development of the Adaptive Guidance and Tips feature. This feature operates in the background, providing unobtrusive, context-sensitive assistance only when patterns of difficulty are detected.
This aligns with her preference for personalized support that respects her growing proficiency
UX Conceptual Design and Framework
From Sketches to Low fidelity wireframes
Design Prototyping and Iteration
CryptoBuddy
Final Screens
Users experience and knowledge about crypto is assessed with a short & fun quiz while on-boarding.
Educational On-boarding
Smart AI assistant provides personalized, real-time, and contextually relevant guidance. The assistant can analyze the user’s current workflow & suggest the next best action
AI Assistant
Transparent AI Permissions
Users are provided with upfront information on the permissions needed to build trust.
Details of the AI assistant
AI Buddy non intrusively offers help when user feels stuck by leveraging the ubiquitous computing principles of invisibility, seamlessness and context-awareness.
AI Buddy recognises multiple incorrect user inputs and offers help while being non intrusive.
AI Buddy Initiating to help without user input
Multiple incorrect user inputs
Users initiates to interact with the AI Buddy when they encounter unfamiliar technical jargon or feel uncertain about how to proceed, enabling them to receive clear, contextual guidance tailored to their needs by leveraging the ubiquitous computing principles of Adaptability.
AI Buddy
Iterating based on user evaluation
In the first iteration, it wasn't clear which permissions are being asked from the user. We added a screen which clearly states which permissions are requested by the AI bot.
Permissions shown up-front
Permissions Information Hidden
The user did not expect to land on the quiz screen, they felt a bit confused. We added a screen to set-up the context before starting the on-boarding quiz
Intermediate screen missing
Quiz introduced up-front
The preliminary findings highlight both strengths and areas for improvement in CryptoBuddy’s design. While the AI assistant effectively supports novice users through real-time feedback and contextual guidance, enhancements in UX writing, user flow and transparency are essential for creating a more intuitive and trustworthy experience. Here are few changes we made basis are user testing:
Reflection
One big challenge we faced was understanding the cryptocurrency space since we were new to it. To tackle this, we started with extensive secondary research, including reading articles and watching videos, to grasp the basics. This helped us build a strong foundation and boosted our confidence in the subject. We then conducted user interviews and market research, which gave us deeper insights and further enhanced our understanding of the space.
We accomplished this project within 2 months, prioritizing swift iteration, regular updates, and actively seeking feedback from potential users, stakeholders & design community.
Future Recommendations
Ensure the AI assistant is more intuitively placed and easily accessible to users throughout their wallet experience.
Develop more sophisticated error detection and prevention mechanisms, particularly for critical actions like transactions.
Use subtle animations or notifications to draw attention to CryptoBuddy, ensuring users are aware of the AI's presence and capabilities.
Relevant Links
Testing Protocol: Data Collection and Testing Plan
Recordings and Observational Notes Folder: CryptoBuddy Test Data and Findings
Testing Methods
We conducted usability testing of CryptoBuddy using a think-aloud protocol complemented by pre and post-surveys. Sessions were conducted both live and remotely via Zoom, recording both participant audio and screen interactions (with consent). Observational notes and survey responses provided rich qualitative and quantitative insights into user behavior and preferences.
Think-Aloud Protocol:
Each session lasted 60 minutes and included interaction with CryptoBuddy’s prototype, focusing on tasks such as onboarding, navigating safety features, and correcting errors during transactions. Sessions were facilitated by a UX researcher, with a note-taker documenting key observations.Based on participants baseline knowledge of cryptocurrency and expectations of CryptoBuddy. Evaluated usability, satisfaction, and knowledge gain using Likert scales and open-ended responses.
Participant Recruitment:
Participants were crypto-curious individuals aged 25–45 with minimal experience. Recruitment challenges were addressed through multiple outreach strategies to ensure diverse perspectives.
Participant Insights
Demographics: Participants ranged from novice to intermediate users with no professional crypto experience but familiarity with digital wallets.
Tasks Completed:
Onboarding: Setting up a wallet, creating a seed phrase, and completing an educational quiz.
Safety Features: Learning to avoid scams and navigating safety resources.
Transactions: Correcting errors and completing a fund transfer.
Data Collection and Findings
Key Findings
Impact on Final Design
Overwhelming Text: Participants found text-heavy explanations challenging and preferred visual aids like animations or tutorials.
Permission Requests: Lack of clear explanations for permissions led to user hesitation.
Subtle Design Cues: Some users overlooked the AI assistant due to insufficient visual emphasis on its role.
Weaknesses
Real-Time Feedback: CryptoBuddy’s AI assistant effectively guided participants during errors and complex tasks like fund transfers.
Educational Features: Step-by-step onboarding simplified technical crypto concepts, fostering user confidence.
Contextual Adaptability: The AI assistant adjusted its interventions based on user needs, aligning with ubiquitous computing principles.
Strengths
These findings informed key prototype improvements, including:
Improving user flow by adding intermediate screens
Enhancing cognitive load by improving UX writing
Providing transparent explanations
for permissions to build trust.
Screens
Loading Screen
Additional Details
AI Helping
Feedback
Setting up AI Buddy
Gathering User Knowledge
Wallet Address Corrected
Transfer Successful
AI Buddy Security Education
Transfer to home page
User Permissions
Understanding user knowledge
User Quiz
Onboarding
Home Screen
Wallet Address Error
AI Buddy Error Recognition
Meet the team
Deepak Mangapuram
Product Designer
Designed high-fidelity mockups.
Created storyboards to visualize user interactions.
Conducted user interviews to gather insights and improve designs.
Conducted usability testing sessions to ensure user-friendly experiences
Sneha Sivasubramanian
UX Researcher
Proposed the project topic.
Conducted secondary research & facilitated user interviews.
Created detailed storyboards & User personas
Conducted usability testing sessions.
Designed high-fidelity prototypes and synthesized findings from research and testing.
Crafted reports to document project progress and outcomes.
Kanika Bansal
Product Designer
My role included literature reviews about different types of cryptocurrency wallets, its uses cases & capability of AI technologies. Additionally, I conducted user interviews with a semi-structured questionnaire to understand user pain points, used the insights to contribute to ideation & built wireframes.
Kavya Maragoni
Product Designer
I identified the problem space through primitive research and conducted in-depth market analysis to uncover gaps in existing crypto wallets. Leveraging secondary research insights, I ideated user-centric solutions and validated them through usability testing with participants. I led think-aloud sessions with participants to identify usability challenges in current crypto wallets, ensuring the proposed solutions addressed real user needs and enhanced trust and usability.