Inspiration
Difficult conversations rarely begin with a lack of words. They begin with confusion.
A manager may be dealing with missed expectations, tension between teammates, poor performance, or a message that suddenly feels confrontational. They may know that something needs to be addressed but still struggle to identify the real issue, decide what outcome they want, and respond without making the situation worse.
Most AI tools provide answers only after the user has already organized the problem into a clear prompt. During emotionally charged situations, however, gaining that clarity is often the hardest part.
I built BubblyCue to support the complete journey—from a tangled situation to a thoughtful plan, and from preparation to the real moment of communication.
Its philosophy is simple:
Notice. Reflect. Respond.
BubblyCue provides another cue, not a verdict. It helps managers think more clearly while leaving the final judgment and decision with them.
What it does
BubblyCue is an AI communication companion for managers navigating difficult workplace conversations.
A manager can describe what happened in their own words, even when the situation still feels unclear. BubblyCue helps them:
- Untangle the situation and identify what is making it difficult.
- Clarify what they want to achieve.
- Organize the important context into a visual Conversation Map.
- Plan a more intentional approach.
- Practice possible responses and receive focused feedback.
- Create a concise Ready Card for the real conversation.
- Receive contextual support during written conversations through a floating bubble.
The floating bubble is a central part of the experience.
With explicit permission from the user, BubblyCue can understand the visible written conversation when the manager requests help. The floating bubble remains accessible over the communication app and provides three practical layers of guidance:
- What I noticed
- Why it matters
- Help me reply
This means the support does not end when preparation is complete. BubblyCue can carry the manager’s situation, intended outcome, and relevant context into the moment when communication is actually happening.
The manager does not need to repeatedly copy, paste, and re-explain the entire conversation inside a separate AI chat.
How we built it
BubblyCue is a native Android application built with Kotlin and Jetpack Compose.
The app uses Gemini to understand communication context, surface relevant cues, help organize complex situations, and generate practical response guidance. Firebase services support the AI integration and application security.
The live-support experience combines user-controlled, permission-based screen sharing with an interactive floating bubble. Users decide when live support begins and when they want BubblyCue to analyze the visible conversation.
The preparation journey is structured around several connected experiences:
- Describing the management situation.
- Receiving an initial cue check.
- Building a Conversation Map.
- Practicing a possible response.
- Receiving focused communication feedback.
- Generating a Ready Card.
- Using the floating bubble when live written support is needed.
Local data storage preserves the user’s preparation and progress, allowing them to leave the journey and return to an active situation later.
Throughout development, I tested the application on a physical Android device, reviewed each result, identified problems, and continuously refined the experience.
I also conducted a Google Play closed test so real Android users could install the application, move through the journey, try its features, and share feedback.
Challenges we ran into
One of the biggest challenges was making BubblyCue feel like one coherent journey rather than another text-heavy AI chatbot.
Real communication problems are rarely linear. A manager may need clarity first, practice later, or immediate support without having enough time to complete the full preparation journey.
The experience therefore needed to support different levels of readiness while keeping the purpose of every screen understandable.
The floating bubble introduced additional challenges involving user permissions, overlay behavior, starting and stopping support safely, maintaining the correct context, and clearly showing when live support is active.
Navigation and state management were also challenging. The Conversation Map, practice feedback, Ready Card, and live-support progress needed to remain connected so the user’s work did not feel fragmented or disappear when they returned to the Home screen.
Another major challenge was balancing useful AI guidance with responsible communication.
BubblyCue should not pretend to know another person’s intentions or present one interpretation as unquestionable truth. Its language is deliberately framed as cues, observations, and possibilities so the manager can reflect and make the final decision.
This principle became part of BubblyCue’s identity:
Another cue, not a verdict.
Accomplishments that we're proud of
I am proud that BubblyCue grew from a floating-bubble concept into a more complete communication journey.
The product now connects:
Confusion → clarity → planning → practice → live support
I am especially proud of creating a working Android floating-support experience that can provide contextual guidance while the user is communicating in another application.
I also built a preparation journey that can preserve an active situation, generate a visual Conversation Map, support practice, provide focused feedback, and produce a Ready Card for the real conversation.
Another important accomplishment was moving BubblyCue into Google Play closed testing. Real testers were able to install the app and provide feedback before its wider release.
Throughout the project, I moved beyond building isolated features and began thinking more deeply as a product designer: testing navigation, listening to feedback, simplifying the journey, improving how the main value is introduced, and separating what works today from the longer-term vision.
What we learned
Building BubblyCue taught me that a technically distinctive feature is not enough if users cannot immediately understand why it matters.
Early feedback showed me that the floating bubble could become hidden behind the preparation journey, even though it is one of the product’s most important capabilities.
That insight led me to rethink the Home experience so users can clearly choose between preparing for a situation and receiving live support immediately.
I also learned that valuable AI communication guidance requires more than generating a polished reply.
The product should first help the user notice what may be happening, understand why a cue could matter, and choose a response that supports their actual goal.
Testing also showed me the importance of clear permission explanations, visible progress, reliable navigation, and allowing users to return to unfinished situations without losing their work.
Most importantly, I learned how much persistence is required to turn an idea into a functioning product: designing, building, debugging, testing permissions, refining the experience, responding to tester feedback, and continuing to iterate when the result does not yet fully match the vision.
What's next for BubblyCue
The next step is to strengthen reliability, improve the first-time user experience, expand testing, and make the transition between preparation and live support even more seamless.
Future development will include:
- Deeper personalization based on the manager’s goals and communication style.
- Support for multiple saved situations with clearer progress tracking.
- Additional practice formats for different workplace scenarios.
- Improved contextual guidance across written communication.
- Stronger privacy controls and transparency.
- Expanded multilingual support.
- Continued improvement based on real user feedback.
- Premium plans offering higher usage limits and deeper preparation tools.
The long-term vision is for BubblyCue to become a trusted communication companion that helps people pause, gain clarity, and respond more intentionally—not only before a difficult conversation, but while navigating it.
Built With
- android
- android-studio
- compose
- datastore
- firebase
- gemini
- googleplay
- gradle
- jetpack
- kotlin
- materiadesign
- room
Log in or sign up for Devpost to join the conversation.