A two-week Prototyping class project was further explored and developed for a five-week independent project in our Designing Innovation course. The initial concept was scrapped after more research. Charlie was the result of our findings.
Problem
My Role
Team lead, research & design
Team partner
Qinyu Ding
Process Overview
We put into practice a combination of research and interaction design principles we learned through lectures, reading About Face and Luma Institute's framework for Human Centered Design; capitalizing on its flexibility and combining methods to gain greater insight at every stage.
Discovery
Our initial timeline for the project was two weeks. We immediately agreed that our first step was to learn more about the disease so, we identified the following questions to answer:
million Americans live with Alzheimer’s
million people will be living with Alzheimer’s by 2050
percent of help comes from family or other unpaid caregivers
Stakeholder Mapping
Based on our preliminary research, we used stakeholder mapping to identify the key people involved and impacted by Alzheimer’s. Our map included the following people:
Interviews
We interviewed six people ranging from their mid-twenties to their late-forties, all working professionals and women. Below, a sample of our questions:
I wasn’t sure what to do or say when she keeps repeating herself. It’s frustrating.
Websites are helpful but overwhelming.
Facebook groups and Alzheimer’s support forums are depressing. Our support group wasn’t a good fit. They were too old.
I don’t have time to listen to everyone’s sob stories. I have my own to deal with; I honestly don't have the time to weed through it.
Synthesis
Ideation
The interviews served as a foundation for our empathy maps which informed the development of our personas. These personas helped guide the direction of content, design requirements and framework — all in an effort to gain clarity regarding user goals and needs.
A persona for an early-stage Alzheimer’s patient was developed to understand what it feels like to be diagnosed with the disease and what challenges are faced. Two additional personas for millennial caregivers were also developed: a physician and a granddaughter. The former was explored to help identify the unique advantage a physician may or may not have as a caregiver. The later persona was developed to understand caregiving in a multi-generational household and one where more than one person develops Alzheimer’s disease.
Discovery + Understanding
In an effort to understand what products and services already exist for caregivers and Alzheimer’s patients, we looked at many types of products and services ranging from websites to physical products. While there were many products for caregivers in general (e.g. medical and monitoring apps) there were very few products that were targeted specifically for caregivers of people with Alzheimer’s. In fact, a review of mobile health apps found that “there is still an imperative demand for resources focusing on supporting the caregiver’s emotional well-being, including coping with stress, anxiety, and depression” (Grossman, Zak & Zelinski, 2018).
Several therapy apps led us to chatbots. Three notable chatbots: Replika, Woebot, and Tess. Replika and Tess are unstructured, highly sophisticated chatbots. Replika is designed to become a user’s best friend over time and Tess is designed to coach people through difficult situations. Woebot, a structured chatbot, uses CBT, lessons, stories, check-ins, and mood-tracking to support peoples’ mental health. A study by Fitzpatrick et al. (2017) proved conversational agents could successfully deliver CBT and significantly reduce depression in people.
References
Fitzpatrick, K. K., Darcy, A., & Vierhile, M. (2017). Delivering cognitive behavior therapy to young adults with symptoms of depression and anxiety using a fully automated conversational agent (Woebot): a randomized controlled trial. JMIR mental health, 4 (2).
Grossman, M. R., Zak, D. K., & Zelinski, E. M. (2018). Mobile Apps for Caregivers of Older Adults: Quantitative Content Analysis. JMIR mHealth and uHealth, 6 (7).
Design
We taught ourselves and discovered the nuance of designing conversations was the most challenging aspect of designing our prototype. We used Amir Shevat’s book, Designing Bots: Creating Conversational Experiences, as our primary source in addition to several posts on Medium and voice interaction documentation from IBM Watson, Google, and Microsoft.
Chatbots require minimal user interface design so the words, the language, meaning, and context takes priority. For this prototype, we referenced language from various therapy chatbots but recognize that with further development, we need a licensed, CBT (Cognitive Behavioral Therapist) who is also knowledgeable about Alzheimer’s and caregivers to assist in the design of the interactions.
Understanding
To understand a caregiver’s experience while using the chatbot, we developed a journey map based on a scenario where the person with Alzheimer’s is repeating themselves. This helped to identify the thoughts and feelings as the caregiver discovers and uses the chatbot for the first time. The journey map narrowed the focus so we could also focus on a sample conversational flow and begin to shape the personality of the chatbot.
First draft of high fidelity mockups.
Prototype
The UI for chatbot design is minimal so we skipped the wireframing process and placed special emphasis on the words, personality, and tone. Our plan included a presentation to get feedback so it was important to present a more polished “Charlie” as the visual design could impact the perception of our chatbot's personality.
Introduction screens to inform and reassure caregivers.
(From Left:Conversation design, caregiving guides initialize prompts in the chat, caregivers can track their conversations as a journal.)
Feedback Session (version 1)
We used Proto.io to create our prototypes. Proto.io was the best option to communicate the look and feel as well as interaction minus full implementation using natural language processing (NLP) and machine learning (ML). We considered creating a prototype made with Twine or using a Wizard of Oz approach but we felt it was critical to provide an experience as close to reality as possible.
This should be the first choice for a caregiver; not a website.
I also like that it is customizable to the user and hypothetically to the users’ unique experience with the Alzheimer’s patient.
I like the layout as if its a text message and I like the three options for aid, particularly option two!
It looks good. The color is calming. But the conversation is at first, a nice pace, then goes too fast.
Conclusion + Future Work
Conversational design is the user experience. This article, Therapy via conversational design by Kathleen Varghese and the documentation for Google Assistant, IBM Watson, and various other sources were critical to learning about the design of chatbots.
Focus on the words and tone to give chatbots the appropriate personality for your target audience. Make accommodations for dead ends and users who wish to “test” the system.
Collaborate with a therapist. Ideally, we will partner with a therapist who has experience working with caregivers; ideally Millennials. We want to make sure the conversations are appropriate, responsible, and sensitive to the nuances and complexity of caregiving.
Additional UX research and usability testing. We would like to do more generative research and eventually design a more formal usability test with representative participants. More data is needed to inform the next iteration and before resources are devoted to development.
Please say hello
LinkedIN — Behance — Instagram — Twitter
Please do not steal. It may bring on bad karma. Thank you.
©2006-2023 All Rights Reserved.