Angel
Be a better human.
Project Overview
Angel is an AI companion app designed to help you reflect, grow, and become a better human through your daily conversations. It listens, transcribes interactions, analyzes emotional tone, flags logical or moral inconsistencies, and even identifies misleading claims. All to support your self-development.
Unlike traditional AI chat tools, Angel represents the shift from reactive assistants to proactive and ambient AI which is always present, quietly working in the background to understand and guide you without interrupting your life.
Role:
Product Designer
Tools:
Figma, miro, claude,
Gpt 4o
Project Date:
2025
Project type:
Personal
Read time:
5 Minute
People engage in daily conversations without reflection or feedback, missing opportunities for growth. Existing AI tools are reactive and intrusive, offering little support for self-awareness.
There’s a need for a private, always-on AI that helps users understand and improve how they communicate without compromising data privacy.
Problem
By combining ethical analysis, NLP, and intuitive design, I aimed to make communication not just easier, but smarter and more self-aware.
Research
Before diving into research, I assumed that most people don’t reflect on how they communicate emotionally, ethically, or factually.
I noticed a growing need for a tool that could help identify miscommunication, emotional misalignment, and especially misinformation or disinformation in everyday conversations.
I also believed that current AI tools were too reactive and lacked the contextual depth needed to offer meaningful feedback after a conversation. These assumptions shaped my research into how people speak, listen, and process truth in real-time dialogue.
Competetive Analysis
To understand the current landscape, I analyzed several AI-powered communication tools such as Otter.ai, Fireflies, Poised, Omi/Bee AI, and Cluely. While these apps excel in specific areas like transcription, meeting summaries, or real-time coaching, they all share a reactive and task-specific approach. None offer a holistic, passive companion experience that blends emotional insight, misinformation detection, and ethical feedback.
Most importantly, none of these tools prioritize long-term personal growth through daily conversation analysis. This highlights a clear gap in the market that Angel aims to fill, offering a proactive, ambient AI experience designed to improve self-awareness and communication at a deeper level.
This map reveals that misinformation and disinformation rank among the top threats in countries like India (1st), the U.S. (6th), and parts of Europe and Latin America (7th–15th), showing its widespread global impact. Nations like Indonesia (18th) and the UK and Mexico (11th) also recognize it as a growing concern. The data underscores a pressing need for tools like Angel that empower individuals to recognize false information in real time—helping to reduce personal, professional, and societal harm.
I began an initial research with friends and family, primarily in the 25-34 age group, to understand perceptions around AI in conversations, openness to feedback, and concerns about trust and overreliance.
High Conversation Frequency
75% of participants engage in 4 or more spoken conversations daily.
Reflects strong potential usage for a tool that passively captures and enhances daily dialogue.
Value Placed on Personal Growth
63% said personal growth suggestions were the most valuable output from such an app.
Emotional awareness was a secondary but still relevant area (25%).
Privacy Still a Major Concern
50% explicitly rejected background listening, even if on-device and encrypted.
Others were open with conditions (security or feature-based).
Universal Self-Reflection
100% of respondents already reflect on conversations, suggesting natural alignment with Angel’s core purpose.
Misinformation Detection is a Key Draw
88% want the app to track and notify them when misinformation is shared in conversation.
This validates one of Angel’s most distinctive features.
User Suggestions
Desired features include:
On-demand control
No ads
Translation
Context-based response suggestions
Strong Interest in Multifunctional
Feedback
38% preferred “all of the above” features (summary, emotional tone, bias correction, moral feedback).
Correction of false or biased statements ranked as valuable (25%).
Openness in Workplace Use
100% would use Angel if it could help improve communication with their boss.
Indicates trust and desire for real-time feedback in professional settings.
Pain Points
Highlighted Pain Points (from User Research)
Lack of Reflection Tools
Users reflect on conversations but have no structured way to review or learn from them.
Difficulty Identifying Bias or Misinformation
Users feel vulnerable to unknowingly accepting false or biased information in casual conversations.
No Feedback Loop in Conversations
There’s no way to improve how one sounds or responds, especially in professional settings.
Privacy Anxiety
Despite interest in features, 50% are uncomfortable with passive listening, even if local-only.
Desire for Context-Aware Guidance
Users want the app to suggest better responses or clarify tone, but only when needed (not intrusive).
User Persona
"Sometimes I wish I had a second brain that could remember and reflect for me."
Tech comfort Level
High – uses multiple productivity
& AI tools
Communication Habits
Frequent meetings, casual chats, and social interactions
AI Attitude
Curious and open to AI, but concerned about privacy and information accuracy
Goals & Motivations:
Wants to grow professionally, communicate more clearly, and avoid missteps in both work and personal life
Pain Points:
Struggles to recall what was said in important conversations, unaware when misinformed, no tool to review his communication
Amir
Age: 29
Location: Austin, TX
Occupation: Marketing specialist
Education: University Graduate
Martial Status: Single
I created this user persona to better understand the needs, habits, and concerns of potential users like Amir, who engage in frequent conversations but lack tools for reflection and correction. This helps ensure that the design of Angel addresses real-world pain points like misinformation, emotional awareness, and privacy sensitivity.
Why did I chose to solve this problem?
I chose to solve this problem because communication shapes our relationships, decisions, and self-perception, yet we rarely reflect on it in real time. I saw an opportunity to use AI not just as a tool for answers, but as a quiet companion that helps people grow by becoming more aware of how they speak, listen, and respond.
Reflect. Refine. Grow.
Ears for
artificial
intelligence
Solution
The main challenge in designing an app like Angel lies in confronting the discomfort around a device listening to private conversations and the growing influence of AI in personal decision-making. While these concerns are valid, it's clear that AI is already deeply integrated into daily life, helping people communicate, work, and live more efficiently.
As we move toward a future shaped by ambient AI, AR, and VR, this shift toward ever-present, context-aware assistants feels not only inevitable, but necessary.
Design Process
For this project, I followed a user-centered and iterative design process, starting with informal research to understand real communication behaviors and attitudes toward AI. Based on insights, I identified key problems, generated feature ideas, and created wireframes and visual prototypes to explore solutions. Throughout the process, I continuously tested and refined the concept to ensure Angel would be intuitive, meaningful, and aligned with real user needs.
Enhancing my crazy 8
sketch with ai
In this phase, I explored eight rapid-fire concepts for how angel could assist users in everyday moments, from correcting misinformation in casual conversations to offering reminders, emotional guidance, and real-time support. Each sketch was created in under a minute to encourage divergent thinking and creative freedom.
After completing the hand-drawn sketches, I used AI image generation to bring these concepts to life, transforming abstract ideas into visually engaging, semi-realistic scenes. This step helped validate the emotional tone and practical application of each concept while bridging imagination with execution.
User Journey
This journey map shows how users discover, learn, use, and benefit from the Angel app. Discovery happens through social media, tech blogs, or word of mouth. Users explore the Angel web or phone app to understand its value. Key actions include real time fact checking, ambient awareness, conversation archiving, sentiment analysis, and speech assistance. These interactions help users make informed decisions, such as reviewing past conversations, detecting intent, and analyzing emotional trends, all leading to clear goals like successful fact checks, communication insights, and self reflection.
Responsive Design
I designed Angel for mobile, desktop, and wearable devices to ensure seamless support across everyday contexts, whether you're on a call, working at a desk, or on the move. Each platform offers a unique interaction style: phones for portability, desktops for depth, and wearables for discreet, real-time cues, ensuring Angel stays accessible and helpful wherever and however you communicate.
Wireframes
These wireframes outline the core layout for Angel across mobile and desktop. The desktop designs focus on dashboards with modules for sentiment, conversations, and insights, using side and top navigation for easy access. The mobile sketches prioritize vertical flow with cards, emotion charts, and real time prompts. Both formats support a responsive, user-focused experience.
Some messy yet genius
hand sketched wireframes.
😉
The home screen is designed for clarity and quick access across both desktop and mobile. On desktop, users benefit from a side navigation bar, a personalized welcome, and modules for mood tracking, misinformation alerts, and toggling Angel’s mic. The mobile layout simplifies navigation with a bottom tab bar, letting users switch between sentiment analysis, the misinformation feed, and speech assistance. Both versions prioritize ease of use and real time access to key insights.
User Flow
This flow shows how Angel supports real-time conversations. After a user begins speaking, Angel checks if it's active. Once activated, it listens passively and determines if a fact-check is needed. If so, it verifies the information, then allows the conversation to continue. The session ends once the conversation concludes, ensuring a seamless, intelligent assistive experience.
Information Architecture
Angel’s information architecture is designed to be intuitive, scalable, and device-optimized. The structure is consistent across phone, desktop, and wearable, ensuring users can easily navigate core features like mood analysis, misinformation detection, and speech assistance.
On mobile, the architecture focuses on simplicity with key sections such as profile, notifications, feeds, and speech assist. Deeper layers include detailed fact checks and shareable sources.
The watch layout highlights quick-access tools like speech assist, mood insights, and misinformation alerts, supporting real time interaction on the go.
For desktop, the architecture expands into a full dashboard with therapy suggestions, privacy controls, and comprehensive settings. The hierarchy supports deep exploration while maintaining ease of access to the most important actions.
This structure keeps Angel focused, responsive, and personalized across platforms.
Initial Prototypes
The early wireframes of Angel focused on core functionality, fact-checking, sentiment tracking, and real-time conversation assist. Built for both desktop and mobile, this prototype helped test how users would navigate between emotional insights, misinformation alerts, and live guidance. Feedback from this stage shaped the layout, clarified hierarchy, and set the foundation for a more responsive and polished experience.
Usability Testing
Custom text here bro
Home Feed variations
Design Thinking
To support Angel’s dynamic and emotionally responsive interface, I chose a set of bold, modern accent colors that work together as a unified visual ecosystem. Rather than relying on a single accent, I used multiple vibrant tones which are red, yellow, orange, and blue to reflect different states like urgency, alertness, calm, and clarity.
This flexible palette enables clear visual signaling across components like misinformation alerts, sentiment graphs, and speech guidance, while maintaining consistency throughout the product. Each color plays a role but never feels isolated and together they create a system that is expressive, balanced, and functional.
Hi Fidelity Mockups & Prototype
After validating the wireframes, I translated them into high-fidelity mockups using a clean and modern UI style. The goal was to make the payment experience feel seamless, trustworthy, and restaurant-friendly. I used a simple color palette, readable typography, and minimalistic icons to guide users effortlessly through the app.
