Design Study About the Experience of Co-creating With AI
School advance project course and internship in Pirr company Mach/june 2021
The aim of this design research project was to further develop the service and mobile application of Pirr from the perspective of user experience. Pirr is a start-up company for erotic literature. The purpose with Pirr is to empower women’s sexual health by enabling them to create, listen and share their own erotic stories. The goal with this project was to create a seamless design where its functions were focused on the users’ needs and pain points to elevate their experience of the service. The project resulted in data collection about potential users, which were visualized as personas and storyboards. It also resulted in a final suggestion of how the design of the application should look and work. This project followed the method of the Design Thinking process where we worked iteratively through the whole design process.
We started off the pre-sprint before starting with the implementational sprints. We called it the planning sprint by investigating research articles that could be relevant towards the user experience of Pirr. Since the service is in an early conceptualization phase, we wanted t o get some inspiration and understanding of topics that could help to create a larger picture of how the different areas of the concept could influence the user experience. We researched the topics of AI and human co-creation, sound properties and synthetic voices, sex tech and digital reading (e-books/audiobooks). This helped us to get a grasp of the complexity and the size of the whole project. The background research guided us to find a focus point and narrow down the subject to a more comprehensible size.
Design Thinking Process
Survey
To be able to empathize with the users and make grounded design decisions, we created a survey in the first sprint. The survey was posted in different relevant groups on Facebook5, LinkedIn6 and Instagram7. We received 102 responses, which provided us with the understanding of the different users’ needs and pain points, which made it possible for us to ground our design decision towards the user group. This step made it possible for us to create personas and storyboards based on the users’ needs and pinpoints. We created the survey and did a pilot survey before posting it to make sure it was working and that the questions were properly done. In the pilot study we noticed that the English jargon we used was perceived to be too advanced for the participants, and we therefore decided to change the language to a more plain language and not too much of an academic or technical language, as recommended by Marsh. The aim of the survey was to look into the users’ preferences, experience with AI and technological habits, experience with erotic literature, reading habits and their willingness and reasons of collaboration with the other people. Therefore, We got an understanding of the three different user groups’ reading habits in and encountering the AI and their expectation of erotic story and Pirr service. This was a big part of the empathize phase to gather empathy for the user and also to be able to define the problem.
Personas
We have created personas to visualize the data and facilitate the communication about the users from the survey. After analyzing the data from the survey, we were able to conceptualize the target groups preferences, pain points and needs in terms of the Pirr service. We identified
three main target groups and created three different personas, which represent the potential users’ challenges in facing Pirr.
The user’s pain point are categorized
as following:
1. Sara is a representation of the user group having hard time in finding interest in new technology
2. Alex is a representation of a user group having hard time to figure out what character would like to be in the story, and they have less creativity in writing the story
3. Johanna is a representation of the user group having a little patient and would not like to waste time
Storyboards
The storyboards are based on the personas which are based on the survey answers. This was done to be able to make design decisions about the application that would suit the users needs and pain points. The storyboards also made it easier to contextualize the service. In the storyboards we visualised how the three personas could use the service of Pirr and how it could solve their current problem and needs. In the storyboard about Sara we show how the persona needs more excitement in her life and personal relationship and gets inspiration from using the application
of Pirr.
LO-FI Prototype
We used sketches, paper and sticky notes as our main design
materials to begin our low-fidelity prototype.
This was done together with the stakeholder to quickly be
able to discuss design ideas, by doing this we could begin to understand each other better in order to generate new ideas
that could help us progress within the project. The artifacts
that we created together with the stakeholder were grounded
out of the personas and storyboards that we had created in the first phase of the project.
Flowchart
In order to visualize the user journey of the Pirr service, we made a flowchart in the ideation phase.
This method assisted us to create the ground structure of navigation for the mobile application and locate needs and pain points of the users.
Wireframing
By finalizing the flowchart we began to create basic
wireframing of the AI co-creation task through sketches and
in Adobe XD
HI-FI Prototype
As a part of this project, we started to create the Hi-Fi prototype after identifying the early problems in the Lo-Fi prototype. The Hi-Fi prototype of Pirr was the clickable and closest resemblance of the final product in terms of its details, functions and UI where it allowed us to conduct the user testing. This true representation of the prototype helped us to gain insight of users behavior while using it and use their feedback and insight to improve the prototype in the iteration process.
Testing
In this phase we conducted user tests to gain insight and gather the users’ feedback in order to improve the prototype.
MODERATED USABILITY TEST - TASK SCENARIOS
To conduct the usability test we did a moderated usability test containing task scenarios (see Appendix 5). We did this because the prototype tested was not fully functioning, and we wanted to steer the participants to test specific functions. We also did this because we used the method Wizard of Oz. We started out following our flowchart and did a manuscript based on that. We also added functions and interaction based on what we wrote in the manuscript. In this way we worked iteratively.
We recorded the user test both on video and with sound to be able to look and listen to it afterwards. The tests were transcribed to make it easier to analyse.
To analyse We divided the answers into categories based on the main pages in the prototype. Then we categorised the answers into different colors, blue notes were from users who tested version B of the prototype first
and pink note tested version A first. This was done to see if there was a difference in the answers depending on which version they tested first. The notes contain quotes from the users or summaries of their answers. We analysed the notes and wrote down interesting ideas or comments that would result in insights.
Insights
The insights from the analysis in the affinity diagram was really helpful for us. Some major insights from the user
tests were that the users had a hard time understanding the landing page with the buttons CREATE, LIBRARY, PROFILE and EXPLORE. They got confused and thought that the buttons were to create a profile and not necessarily to create a story.
We found out that a lot of users wanted to be able to find their own work in the PROFILE page and maybe don’t have two different pages for PROFILE and LIBRARY. A lot of the users also said that they would prefer to start in the EXPLORE page when they are new users of the application, since they want to see how other users have created a story. They want to explore other stories before creating one themselves. They also had a hard time understanding how to create a story and wanted onboarding when they were to create a story for the first time. 50% of the users liked version A better with the options of DO, SAY, STORY and 50% of the users liked version B better with different options. However, most of the users preferred when the AI continued the story from the sentence or prompt they wrote as in version A, the participants said it felt more personal and like they had more control of the story compared to in version B when the AI integrated the prompt in a text. The users also preferred to interact with the AI with voice interaction because it felt more personal and like they were having a “real” conversation with the AI.where the test persons thought of the AI in Pirr as more humanlike and having more personal characteristics. This is maybe something Pirr could work on and develop in the future. Another insight was that some of the users thought that the sound editing was placed too late in the creating process and that it should be more integrated with the creating part. They wanted to be able to create the story and add sound at the same time to make the creating process easier and more smooth.
