A chatbot app for self-care during

long-term unemployment.


Amelio is a chatbot app that personifies a trusted source of information and guidance for users going through long-term unemployment (LTU). Defined as being unemployed for 6 or more months, LTU negatively affects individuals’ mental health and physical health, in addition to their financial stability.

This companion addresses needs regarding emotional well-being, financial stress, and lack of motivation, helping users to practice better self-care, budget their spending, set goals, and manage their unemployment benefits.

For this personal project, I took on the roles of UX Researcher, UX Designer, and UI Designer. Using a human-centered design approach, I was responsible for the entire process including user research, ideation, prototyping, and usability testing.

Role: UX Researcher, UX Designer, UI Designer

Date: December 2019 - July 2020

Tools: Sketch, InVision, MarvelPOP, Miro, Google Forms

Key Skills: Human-Centered Design, Screener Surveys, User Interviews, Affinity Maps, Empathy Maps, User Stories, Sketching, User Flows, Wireflows, Wireframing, Style Guide, UI Design, Usability Testing, Interactive Prototyping

Design Process

1. Empathize

  • Understand the Problem

  • Secondary Research

  • Competitive Analysis

  • User Research

2. Define

  • Affinity Mapping

  • Empathy Mapping

  • User Personas

  • Reframe and create human-centric problem statements

3. Ideate

  • User Stories

  • Concept Generation

  • Sitemap

  • User Flows

  • Sketching & Wireframing

4. Prototype

  • Moodboard

  • Style Guide

  • High Fidelity Prototyping

5. Test

  • Usability Testing

  • Audit Report

6. Evaluate

  • Evaluate Feedback

  • Identify Issues and Opportunities

  • Integrate feedback into prototype, re-test, and re-assess

1.  Empathize

Understanding the Problem

Losing your job can be a stressful experience, and the journey that follows is likely to be an increasingly confusing and challenging path. I know this to be true—I was inspired to explore this problem based on my own frustrations dealing with long-term unemployment. Through repeated visits to the Department of Labor, shuffling from waiting room to waiting room, meeting to meeting, and observing people around me from all walks of life, somber and also frustrated, it was obvious to me that there were a lot of opportunities for improvement. I wanted to dive a little deeper and discover what kinds of tools, guidance, and resources could exist beyond these walls, that could help make this journey more manageable. I wanted to understand the bigger picture and reimagine new possibilities.

Secondary Research

To gain more insight into the context surrounding the state of unemployment as well as its effects on people's lives in the US, I conducted secondary research by looking into news articles, academic publications, books, and government data. 

When I first started research for this project in December 2019, the unemployment rate in the US was at 3.5%, or 5.8 million unemployed people.

For these individuals, as time passes, there is even more at risk besides loss of income. Research shows that long-term unemployment (LTU), defined as being without a job for at least 6 months, has a lasting impact that spans beyond financial insecurity—it negatively affects individuals’ mental health and physical health as well.

When people lose their jobs, they also lose values such as a structured day, shared experiences, and status. This is a loss that consequently affects self-esteem, leading to anxiety, self-doubt, and depression, and adds to the already difficult and stressful experience of looking for work.

Fast-forward a few months, with the devastating emergence of the global COVID-19 pandemic, unemployment now affects 30 million people in the US—that's about 1 in 5 American workers.

With this data in mind, long-term unemployment is likely to become a growing socio-economic issue with lasting effects that more and more people will struggle with over time.

Competitive Analysis

Now that I had a better understanding of the problem space, I wanted to learn more about the kinds of resources that already exist for people who are unemployed long-term. Where can users find the support they need? What are their options?

I started my competitive research by looking into government resources and unemployment benefits first. I decided to avoid job search websites because, while there is definite value in those experiences for those who are unemployed, I wanted to explore gaps and opportunities that may be present beyond those tools. 

Competitive Analysis chart comparing various existing resources and their ratings

I identified and evaluated seven competitors in the problem space of unemployment: the NYS Department of Labor Portal, Benefits.gov, CareerOneStop, MySkillsMyFuture, FileUnemployment, FreshEBT, and Wisdo. Looking at several heuristics for interaction design, I noticed that there was a wide range in usability issues as well as successes.

While competitors generally rated well in terms of aesthetics and help and documentation, many were lacking in other areas of usability including visibility of system status and flexibility and efficiency of use. These were heuristics that I assumed to be some of the most important things to users when dealing with unemployment resources.

Understanding the User

Despite having been unemployed myself, what might be true for me might not apply to other end users of this project. Prioritizing a human-centered design process, I was eager to learn about others' experiences in this problem space. 

So, who are the end users I'm trying to learn about? To answer this question and hone in on their needs, I was eager to talk to people and gather some primary research of my own. I started with screener surveys for initial quantitative research, and from this pool of participants, I recruited 5 individuals for qualitative research through 1:1 interviews.

Screener Survey

Through a quick online survey, I was able to gather initial insights from 43 potential users, all of whom had been unemployed for a minimum of 3 months within the past 5 years, and use web/mobile apps regularly. Through their responses, I was able to better understand my users’ experiences when dealing with unemployment.




Age Range


have been unemployed for 6+ months


felt discouraged, depressed, or unmotivated while unemployed

User Interviews

I used the results from the screener survey as the basis for an open-ended interview script, and then chose 5 potential users who met specific criteria of my target user: someone who has been unemployed for 6+ months and has struggled with mental health as a result of their circumstances. 

My goals for this qualitative research method were to:

  • Identify pain points, goals, and motivations for people going through unemployment

  • Understand user behavior when dealing with unemployment

  • Validate my initial research of the problem space ​

After conducting the interviews and transcribing each recording, I was ready to analyze my qualitative data, in order to further define the problem and the user as well.

2.  Define

Post-Its were grouped by category, with each color representing a different interviewee

Primary Research Insights

  • Common experiences of rejection, self-doubt, imposter syndrome

  • More negative feelings when people didn’t have plans after they left/lost their jobs (uncertainty), more positive feelings when they did have plans

  • Generally, people had a difficulty getting unemployment/health benefits, but they needed it—finances are a challenge

  • Sense of accomplishment from small steps or actions

  • Reaching out to friends, family, mentors, etc. was helpful

Capstone - Affinity Map.png

Affinity Mapping

With my interviewee's responses written down on dozens of Post-Its, I used affinity mapping to help me categorize my notes, identify patterns, and distill insights.

Empathy Mapping

Insights gathered from affinity mapping allowed me to then create empathy maps, in order to further organize my observations and understand potential users' collective behaviors and needs.

Empathy Map - Persona A.png

Example of an Empathy Map representing one of my two personas

User Personas

With clarity from the empathy maps, I created user personas to define two distinct potential users.

Persona 1 - Olivia.png

Olivia, the Professional

  • Has been unemployed for 6 months and is trying to find a job she will love

  • Receives unemployment benefits and is looking for more financial assistance

  • Wants to start budgeting, stay focused, and maintain her mental health while dealing with rejection

Ben, the New Grad

  • Recent college graduate looking for his first full-time job, and is learning along the way

  • Does not qualify for unemployment benefits, but is struggling financially and emotionally

  • Looking for guidance to stay focused on his job search, save money, and regain some confidence

Persona 2 - Ben.png

How Might We...?

My analysis led to me to these 3 How Might We statements to help further define the problem of long-term unemployment:

  1. How might we ease feelings of rejection, self-doubt, and imposter syndrome for people who have been unemployed for 6+ months?

  2. How might we assist unemployed individuals with planning out their work search goals and to-do’s, now that they no longer have regular schedules?

  3. How might we offer guidance for unemployed individuals to learn about and apply for the benefits they need?

3.  Ideate

User Stories

In creating user stories, I started with themes, to broadly categorize users’ needs: Motivation, Mental Health, Work Search, and Finances. From there, I created epics that my personas Olivia and Ben would have as goals, and then specified tasks and benefits that they would most likely prefer. From there, I rated each story by priority, and narrowed down the list further by selecting 2 of the most important stories from each category to serve as the basis for the MVP.

User Stories organized by themes

User Flow

To help me understand users' most critical routes and initial information architecture, I created a User Flow diagram. With this tool, I mapped out three main paths that users would take within the user experience, based on the high-priority user stories that I selected above:

  1. Practice self-care through meditation

  2. Setting a goal

  3. Setting a budget

User Flow diagram mapping out the most critical routes within the app

Sketches and Wireframes



With the User Flow in mind, I sketched out critical screens to get a tangible sense of how users might follow critical routes. Using these sketches, I created an interactive paper prototype for usability testing.

Preliminary sketches of select screens


Next, I created ​low fidelity wireframes to clarify the direction of interaction design and address any initial usability concerns. This became my basic outline for the overall design.


Wireframes of select screens

4.  Prototype



Taking into consideration the stressful nature of unemployment, I wanted the visual design to feel uplifting, encouraging, trustworthy, and calming.

I was inspired by soothing cool colors with pops of warmth. I also took inspiration from existing mental health apps to help me visualize what the user interface could look like.

Visual Design

Based on my moodboard explorations, I selected a color palette based on teal as the primary color, with a light teal and complementary warm colors to round out the palette, and made sure that these colors satisfied accessibility considerations. To communicate a more friendly and approachable user interface, I decided on rounded buttons and chose a modern sans serif font with rounded letters.

I incorporated these elements into an evolving style guide, which I updated throughout prototyping and testing to keep the user interface visually consistent.


Hi-fidelity Prototypes

Using my moodboard and style guide to direct the visual design of the project, I was ready to revisit my wireframes and create a high fidelity prototype to refine the app.

Version 1.0

In my first iteration of the hi-fidelity prototype, I included more personality and created a character to embody the conversational user interface of the app. While working through the screens, I was able to choose a name for this character and app as well: Amelio. 

A small selection of my first round of hi-fidelity screens

Version 2.0

After discovering various usability and heuristic issues in the first version of the prototype, I refined my screens further to help solve those issues. A few of the biggest changes I made were to the conversational user interface with Amelio, the budgeting wizard, and the goals wizard—overall fine-tuning options to allow for more simplicity, flexibility, and control for users.

5.  Test

Usability Testing

As I was prototyping, I always incorporated usability testing into my process before creating the next iteration. This way, I was able to gather qualitative data on first impressions, usability, and opportunities for improvement to inform any necessary changes to my designs. I learned about what worked and what didn't work, very quickly. 


Throughout the process, I was able to test with 5 potential users each time I created an interactive prototype, for a total of 15 participants over 3 rounds of testing. While testing my high fidelity prototypes, I synthesized my findings by categorizing all issues and comments participants made about each prototype and gave each one a severity scale rating based on how critical the issue was. I tracked all concerns in a spreadsheet to allow for a more structured way to distill insights and priorities. 

Test Results for Version 1.0

  • Users need to land on a results page after completing the Budget Wizard to validate their efforts.

  • Users do not know if they can multi-select certain buttons within the wizard UI.

  • Users would like more control and clarity within the Budget Wizard, particularly for privacy and flexibility.

Test Results for Version 2.0

  • Users need clarity for functionality of radio buttons within the Goals screen.

  • Users need more control over sorting and filtering their goals.

  • Users would like to see a history of their activity within the Budget screen.


Through multiple rounds of usability testing with 15 different potential users, some of the biggest key learnings I gained were:

  • Forms need to balance simplicity with flexibility, so that users don't feel overwhelmed but still have control if they want to customize their options.

  • Chatbots present a level of uncertainty to users, so users like seeing examples of how they can best use them.

  • Users value visual clarity and validation, especially if they are expected to input any personal information.

  • Even if the design is well-meaning, users will have concerns about privacy that may affect their perception of the product.

  • Incorporating mental health, goal-setting, and budgeting all into a single, cohesive app is unfamiliar territory, for myself and for my users. Tying these features together in a way that was clear and easy-to-use was the biggest challenge.

Closing Statement

Many of the potential users I interviewed expressed that they would use an app like Amelio as use a tool for balancing self-care, goal-setting, and budgeting with unemployed. While there are still critical usability issues to solve, for people dealing with long-term unemployment, I know that I am on the right track.

My next steps for this project will be to incorporate users' feedback into my next iteration, to further refine the app. In addition, since I have only created the most critical screens, I plan to create more screens to complete secondary user flows, to provide more context for the user experience of the app. Examples of these include exploring the meditation library and applying for unemployment benefits.

As unemployment continues to be complex issue in the US, especially in the wake of COVID-19, there is also an opportunity to re-evaluate what users' most crucial needs are during this time, and how Amelio may be able to adapt to those needs as it evolves beyond a MVP.

6.  Evaluate


  • LinkedIn

©2020 by Cynthia Poon.