Introduction - Arizona Water Chatbot
Leveraging AI, the Arizona Water Chatbot provides Arizonians with critical insights on water conservation and management. This project details the application of UX research methodologies and processes to refine and enhance the chatbot's development.
UX Research Methodologies Followed
Heuristic Evaluation
Comparative Analysis
User Interviews
Coding Interviews
Usability Testing
Data Analysis
Chat GPT vs Google Bard (HE)
Comparative Analysis
User Interviews
In the "User Interviews" section, we explored how individuals currently seek out information on water-related issues and gauged interest in utilizing a chatbot for this purpose. We aimed to understand their preferences and assess the potential impact of integrating a chatbot solution into their information-gathering process.
Google's Strength: Broad Information Access
Users appreciated Google for its efficiency in providing a broad overview of water-related information. They found it reliable for general queries, highlighting its role as a valuable starting point.
Identified Gap: Locating Specific Local Information
When seeking detailed information specific to their locality, such as updates from the city's water department, users encountered difficulties. This gap points to the potential benefit of a dedicated tool that simplifies access to localized water information.
Coding Interview Responses
The data was analyzed using a bottom up approach where we organized our findings into groups and then labeled them. We analyzed the “think-aloud” responses that we received from all the participants.
The 'faceless' Persona
This section introduces a 'faceless' persona to guide the chatbot’s development, focusing on user behaviors and needs while avoiding stereotypes. It provides an overview of the users' motivations, goals, technical abilities, and challenges in finding water-related information, informing the chatbot's design to better meet such users' needs.
Usability Testing
A usability test was conducted to evaluate the performance and user experience of the Arizona Water Chatbot. This testing aimed to identify how effectively the chatbot provided information on water-related queries, its ease of use, and the overall satisfaction of users interacting with it. Through this process, we sought insights into the chatbot’s intuitive design, response accuracy, and areas for improvement to ensure it meets the informational needs and expectations of Arizonians.
Setting up Usability Test Materials
Defining Roles
Consent Form
Scenarios with Task
Pre and Post Task Questionnaire
Water Chatbot - Usability Test Results
Using the bottom-up approach the participant responses were categorized with the following codes. It helped in determining the area of user frustrations and delight.
Positive Responses
04
Negative Responses
10
Recommendations
12
Open AI - Custom GPTs
For the next stage of the project, I proceeded to develop a custom GPT with the same instructions given to the already existing water chatbot and the responses were compared.
View comparison analysis
🚧 This project is live and being refined, comeback to view further developments! 🚧
