InPower

InPower is an alternative social media app for those who have suffered abuse, harassment and cyberbullying and are looking to be part of a safe space online. Currently in its beta phase, InPower is actively working to add new features as it builds a community of empowered users and continues to grow.

As part of the product design team, I worked to design a chatbot feature that is powered by AI and trained by psychologists and experts to provide users with personalized support, as well as access to a variety of mental health resources.

Scope

UX Research
Strategy
Information Architecture
UI Design
Wireframing
Prototyping
Testing & Iterations

Client

InPower

Role

Product Designer

Duration

5 Weeks

Creating a safe space in a digital world.

The Challenge

Despite being more connected than ever in today’s digital world, the power of social media has its consequences. Cyberbullying and online harassment are more prevalent than ever and have been proven to have profound, lasting impacts on self-esteem and mental health.

The Opportunity

InPower is working to create positive change. Unlike traditional social media platforms, InPower is built on a foundation of kindness, empathy and personal growth and aims to empowers users to become their best selves in a safe environment.

With InPower actively adding new features, my team and I set out to design InPower’s AI-based chatbot, which provides users with a safe, anonymous space to talk through their feelings and seek personalized support.

The Process

Empathize - UX Research

InPower’s overall target market aims to eventually be inclusive of all users, but currently they are focusing on young girls, women and non-binary individuals, ages 13+. I wanted to better understand this target market, specifically in reference to their feelings surrounding chatbots and AI.

  • Understand how much knowledge users have about chatbots and AI.

  • Determine the reasons behind why users utilize chatbots and AI.

  • Understand how users implement chatbots and AI in everyday life.

  • Discover what features users enjoy or expect when using chatbots and AI.

  • Learn what users like and dislike about the use of chatbots and AI.

Competitive Analysis

Before delving further into understanding our users, I first wanted to get a better grasp on the current market of competitors and their usage of chatbots and AI. Conducting a competitor analysis afforded me with valuable insights and allowed me to investigate industry trends, visualize user expectations and identify feature gaps. I explored general AI-based chatbot competitors, as well as those more focused on user mental health.

Insights

  • There are varying degrees of AI intelligence amongst chatbots, with some offering more human-like, personalized experiences.

  • Additional chatbot features such as mood tracking, goals, interactive tools and more are commonly used to increase user engagement.

  • Many chatbot competitors provide tailored and customizable experiences for users.

User Interviews

In order to better understand how users feel about chatbots and AI, I conducted user interviews with the following research objectives:

Key Findings

While the majority of participants mainly utilize chatbots and AI as a tool to help with daily tasks or for educational purposes, they could see the value in a chatbot that can show empathy and provide support as a mental health tool. Participants also like the idea of a chatbot that can learn from their interactions and cater to their individual personalities and needs, however they distrust chatbots that are too personified and wish to be clear on boundaries. Users are also more likely to trust the advice of the chatbot if they know that it is trained by licensed therapists and psychologists.

User Pain Points

Participants were unanimous in their desire for privacy and data security policies to be fully transparent. They expressed hesitation about disclosing sensitive or personal information to a chatbot if their weren’t strict privacy regulations in place. Participants also expressed a general fear of becoming too reliant on AI both as individuals and as a society.

Actionable Insights

Now that I have a better understanding of my users, I can focus on the following actionable insights to help inform my design decisions moving forward:

  • Prioritize privacy & data transparency so users are fully aware if their personal information is being used.

  • Offer user anonymity so that users will feel more comfortable disclosing sensitive information.

  • Build user trust with AI that is trained by psychologists and experts to offer support and guidance.

  • Personalize the experience with a chatbot that learns from user interactions.

Challenges

One of the biggest challenges I faced with this project was having to set aside my assumptions and stay neutral despite my own personal biases surrounding AI. It was important to stay neutral and unbiased in order to fully learn and understand my users.

Define - User Empathy, POVs & HMWs

Building User Empathy

My competitive analysis and user interviews provided me with a better understanding of who my users are and what problems they face. In an effort to turn my research findings into concrete visualizations, I created two user personas to help me empathize with my users and understand their needs, motivations and pain points.

POVs & HMWs

With my UX research and user personas helping me to better understand, empathize and visualize my users, I wanted to create point-of-view statements that reflect the problems and needs of InPower users. By considering the issues that users face, I can then focus on how-might-we questions that will help me think of solutions to user pain points and ways to meet user needs.

Problem 1:

POV: Chase loves the idea of talking through issues with a chatbot, but their fears of private conversations being monitored and their data being leaked make it hard to truly open up.

HMW:

  1. How might we build trust with our users so that they can feel secure when it comes to privacy and data in order for them to truly open up?

  2. How might we design a chatbot that properly informs users about the security of their privacy and data?

  3. How might we create a safe space that protects our users so that they have nothing to fear?

Problem 2:

POV: After a negative experience left Sylvia afraid to share her personal thoughts with others, Sylvia would love a safe space to be able to express her emotions and thoughts without the fear of judgment.

HMW:

  1. How might we create a safe space for users to express their emotions and thoughts without the fear of judgment?

  2. How might we provide the right tools to give our users the freedom to share their feelings?

  3. How might we assure our users that they won’t be judged so that they can truly feel secure when expressing themselves?

Problem 3:

POV: As someone who suffers from social anxiety, Quinn struggles with initiating and maintaining conversations and as a consequence never feels that they can truly connect with others.

HMW:

  1. How might we empower individuals with social anxiety to build confidence in initiating and maintaining conversations so that they can forge real connections?

  2. How might we provide resources that could be beneficial in helping users with social anxiety?

  3. How might we help users navigate their problems so that they can learn and grow?

The Big Question:

How might we build trust with our users and create a safe space to share thoughts and feelings without the fear of judgment so that they may foster healthy connections, learn and grow?

Ideate - Feature Roadmap, Sitemap & User Flow

Turning Needs into Features

With my POV and HMW Questions helping me to better understand the problems and needs of my users, it was time to turn those user needs into product features. Keeping my research and user personas in mind, I created a product feature roadmap which ranks features from high to low priority. Each feature was thoughtfully designed to enhance the user experience, while meeting user expectations and goals. View below or access the feature roadmap here.

Establishing Hierarchy

In addition to my product feature roadmap, I wanted a way to visualize the organization and hierarchy of the chatbot features. Being an added feature on the InPower app, I created a sitemap specifically for the chatbot as a stand-alone feature.

Crafting an Intuitive Experience

In order to craft an intuitive experience for InPower users, I wanted to create a user flow that illustrates the journey users take when interacting with the chatbot. From launching the bot to engaging in conversation, the user flow maps out user interactions and highlights potential pathways in order to ensure a seamless user experience.

Prototype - Wireframing, Branding & UI Design

Low-Fidelity Wireframing

With my user flow mapped out, it was time to put pencil to paper. I started by sketching out various ideas and concepts, allowing myself the freedom to think outside the box without criticizing any one idea just yet. As a designer, I always start with sketching first before moving to Figma to create low-fidelity wireframes of my favorite ideas.

Branding & UI Design

What’s in a Name?

While InPower already has established branding in place, my team and I were tasked with developing the chatbot UI, including the chatbot name and icon. Based on our user research, we knew that users distrust chatbots that are too personified and dislike AI that mimics humans too closely. Keeping this in mind, we knew that our chatbot icon needed to be uniquely bot-like, while still being friendly and approachable.

While brainstorming icons, we also explored potential names for our chatbot. During a meeting with our stakeholders, InPower’s founder and CEO expressed wanting a name that was both unique and meaningful. She loved the idea of naming the bot after her friend and inspiration, Dominique, but struggled to find a way to adapt it into a fitting name for the chatbot.

My team was left to do some thinking. We wanted a name that was cute and fun, especially for InPower’s younger target market, but also wished to find a way to make our client happy. While researching alternative names for Dominique, we discovered the nickname ‘Meeka.’ It was cute, friendly and uniquely meaningful to our client. Thus Meeka was born!

Meet Meeka!

High-Fidelity Wireframing

With Meeka starting to evolve, my team and I began the process of high-fidelity wireframing. We created the main chat flow, and then focused our attention on various interactions users could experience within the chatbot. My focus was on creating the activities icon pop-up, as well as the daily check-in flow. From there, I was able to prototype and prepare for usability testing. Take a sneak peek below or view the high-fidelity wireframes here.

Test - Usability Testing & Iterations

Usability Testing

With my prototype ready, it was time to test. Each member of my team conducted usability testing with five participants in an effort to evaluate how effectively users are able to navigate the chatbot UI and engage with our key interactions.

Success Metrics:

  • Task success rate

  • Time to complete tasks

  • Ease of use

  • Error rates and issues encountered

  • User satisfaction (scale from 1-5)

Methodology:

The usability testing platform Lyssna was used to host the tests and interviews were conducted both in-person and virtually via Zoom.

Users were asked to independently perform each task to the best of their abilities while I observed their interactions. After each task, I engaged them in conversation to discuss their overall thoughts and any frustrations.

Task 1: You just learned about InPower’s new chatbot feature and are eager to try it out. Find the chatbot and initiate a conversation.

Metrics:

  •  100% reached the goal screen

  • Average time spent: 93s

  • 3% misclicks

  • 100% of users found this task to be very easy.

  • 100% of users felt very satisfied using this product.

Insights:

  • All users found the welcome page, privacy notice and chat very easy to navigate through.

  • While all users were able to locate the chatbot icon, some did mention that the lighter color was harder to notice floating on the homepage.

  • Most users found the design intuitive, but there was some confusion on whether Meeka was a bot or access to chat with a live therapist.

Task 2: You are feeling anxious. Talk to Meeka and complete the Daily Check-In.

Metrics:

  • 83% reached the goal screen. One user had technical difficulties.

  • Average time spent: 91s

  • 41% misclicks

  • 100% of users who completed this this task found it to be easy to very easy.

  • 100% of users who completed this task felt very satisfied using this product.

Insights:

  • The majority of users found this task easy to navigate without many challenges.

  • Most users loved the emojis for emotions and the mood slider.

  • Users appreciated the dynamic highlighting of buttons.

  • Some users didn’t realize right away that the emotions were scrollable.

  • Some users tried to click into the text box before selecting the emotion.

Key Findings:

  • Users found the overall design intuitive and easy to navigate.

  • The Meeka icon received positive feedback, but the colors needed some attention in terms of accessibility.

  • Greater clarity is needed to ensure users fully understand Meeka’s purpose and the services it offers, eliminating any potential confusion.

  • Clearer hierarchy is necessary on the Daily Check-In process.

  • Users enjoyed the animation and emoji usage for the mood slider, but didn’t intuitively recognize the ability to scroll through the emotions so this requires further attention.

“Meeka was very intuitive to find! Very unique icon, welcoming, friendly, un-intimidating and inviting.”

“Nice to have emojis for emotions! Dynamic highlighting of buttons made the experience more intuitive.”

“At first I didn't realize you could scroll with the mood slider but I love the animation and variety of moods.”

Iterations

With the help of my usability test results, I identified three key areas for improvement:

Accessibility

Usability testing showed that on the InPower homepage, users found the floating Meeka icon hard to notice at first. For the sake of accessibility, I decided to move the Meeka icon to be included in the top navigation menu so users can more easily access the chatbot.

Enhancing Clarity

On the welcome screen, adding in the text “AI friend” helps clarify that Meeka is indeed a bot, and not an actual therapist. While users, can book therapy sessions with real, licensed therapist through InPower and with Meeka’s help, our usability testing showed that it was important to clarify that Meeka is an AI bot.

Refining Hierarchy

Usability testing indicated that users did not always realize that the mood slider was scrollable. To correct this, the sizing of the mood slider was changed to be slightly smaller so that a bit of the emojis on the sides show, making it more visible and indicating to the user that there is more to see. Additionally, greying out the text box until users select an emotion will help users follow the design hierarchy and avoid confusion.

The Final Prototype

After completing iterations and presenting the finalized product to our stakeholders, the chatbot is finally ready to be integrated into the InPower app! Check out the high-fidelity prototype below to see what Meeka can do. Be sure to keep an eye out for the pink activities icon which can expand and will eventually link to future design screens and activities for users within the app. Let’s meet Meeka!

Reflection - Final Thoughts, Challenges & Next Steps

Final Thoughts:

As my first client-based team project, working on InPower’s chatbot feature has been an invaluable learning experience that has helped me to grow immensely, both as a designer and individual. InPower’s mission is close to my heart and I am so proud to have been able to be a part of their journey to create a digital safe haven. From getting to know and trust my teammates and learning about our users, to bonding with the clients and efficiently collaborating across other departments, my self confidence grew as Meeka took shape. I’m incredibly proud of our little AI powered chatbot and can’t wait to see all the good it will do!

Challenges:

One of the biggest challenges I faced while working on this project was learning to defend my design decisions against stakeholder wishes. As a UX designer, there is always a balance of meeting client requirements while prioritizing user needs. Learning to trust my user research and find my voice as a UX designer was challenging, yet incredibly rewarding. Putting this into practice helped my confidence to grow and was perhaps where I experienced the most growth as a UX designer throughout this project.

Next Steps:

Currently, InPower is actively developing their platform with plans of launching to the public soon. I’m excited to stay connected with the team and look forward to following their journeys. As for Meeka, there are several opportunities for enhancement. If I were to continue working on the chatbot, I’d start by expanding the activity icon’s functionality, adding more interactive features such as guided breathing exercises in an effort to boost user engagement. Additionally, I would love to see Meeka become fully customizable, allowing users to tailor their experience. With AI’s boundless potential, I believe Meeka could significantly enrich the user experience within the InPower platform and I’m excited to see what’s next!

Thank You!