Scheduling meetings in a business environment can be a time-consuming and cumbersome process, involving multiple emails or phone calls to find a mutually available time. This can result in frustration, missed opportunities, and wasted time for busy professionals. Some of the main problems that we encountered while conducting a survey study were:

When I first joined the team, we already had a design prototype, but it had not been validated. We were uncertain about whether the design would meet the needs of our users, and we needed to conduct interviews to gather feedback. Given that Organ.AI is an AI app that requires AI interaction and response, I decided to spend time implementing the prototype directly. This allowed us to quickly validate the solution and make any necessary improvements.

As a product designer, my first step upon joining the team was to quickly build a proof of concept for Organ.AI and validate the idea with potential users. To do this, I leveraged Xcode and some existing libraries and components to build the app rapidly. Through rapid prototyping and user testing methods, I gathered valuable insights and feedback that allowed us to iterate on the design and development of the final product, ensuring that it met the needs of our target users.

"During the first round of testing, we conducted a beta-loop test where we asked 12 participants to connect their own calendars and use the app for a week. We collected feedback from them and identified several key issues, including a lack of feedback, unclear instructions for getting started, and time-consuming data input due to typing.

Following the first round of testing, we held a workshop with the development team to prioritize the issues identified by the users and generate ideas for potential solutions. By working collaboratively and drawing on the feedback from the beta test, we were able to identify key features and design changes that would improve the user experience and address the concerns raised by our initial testers.

We generated two potential solutions during the workshop: a text-based chatbot and a button-based chatbot. Rather than spending time on building a prototype, we decided to perform A/B testing with existing tools in the market to determine which option was more effective. We utilized a messaging library to create the text-based chatbot and Landbot.io for the button-based chatbot, both of which were integrated with our AI backend to fully support our users in scheduling meetings.

Continuing the case study, to gather insights on the effectiveness of the two chatbot designs, I conducted a usability testing with 9 participants and collected their feedback. Additionally, we used the System Usability Scale (SUS) questionnaire to measure the overall usability score and timed how long it took for participants to complete tasks using each design. The data from this testing helped us to make an informed decision on which version to move forward with.

Based on feedback from user testing and research, we made several key improvements to our scheduling assistant app. These included integrating button-based interactions into our text-based prototype, adding an auto-complete feature to minimize typing effort, and including more personalities in our AI.

After making improvements to the design based on feedback from previous usability testing, we conducted a second round of testing with participants using the improved chatbots. Our goal was to measure the effectiveness and efficiency of the changes we made and gather further insights to inform the final design of our AI scheduling assistant.

After addressing the feedback we received, we conducted a final polish on the design to ensure consistency and improve the overall user experience. With the design now finalized, we handed it off to the development team for implementation.