In this subject, we carried out a project that involved developing a simulated bot called "Be Stated" to assist students from the Faculty of Mathematics at UADY in practicing their English and reinforcing their language confidence. Through three deliveries, we identified a social issue, conducted surveys to understand the users' needs, and developed a prototype implemented on the Alexa voice platform.
Although we couldn't conduct usability tests with real users due to time constraints, we managed to draw relevant conclusions based on our analysis and knowledge of the bot. We recognized the importance of properly planning time and allocating sufficient resources for all stages of the project.
We learned valuable lessons throughout the process, such as the importance of confidence when speaking a new language and the need to consider usability tests with real users to obtain direct feedback. We also acquired additional skills, such as conducting surveys and data analysis, and strengthened our abilities in prototype design and usability testing.
During the project development, we encountered challenges related to the lack of consideration for voice design guidelines, especially on the Alexa platform. Our initial idea was to allow the bot to generate completely random conversations; however, we realized that the Chat GPT API didn't efficiently allow storing the context of previous responses. While it was possible to implement it, it wasn't cost-effective due to the limited infrastructure we had and the high cost in terms of the number of tokens required.
This setback taught us the importance of considering the limitations and specific design guidelines for the platform being used. We learned that it is essential to understand the capabilities and constraints of the tools and technologies we employ in order to design effective and efficient solutions.
Despite this obstacle, we managed to find a viable solution that allowed us to proceed with the bot's development and meet the project's objectives. We learned the importance of adapting to the resources and constraints at our disposal and seeking creative alternatives when facing technical limitations.
In future implementations, we would carefully consider voice design guidelines and evaluate the available options for storing and managing the context of previous responses more efficiently. This would enable us to improve the quality and coherence of the conversations generated by the bot.
In summary, this project allowed us to apply theoretical knowledge in a practical manner, which helped identify errors, areas for improvement, and promote subject knowledge. Although the bot was simulated, we believe that a well-implemented version could serve to reinforce students' confidence in the English language. We are proud of the results achieved and excited about future opportunities to apply our skills in challenging and meaningful projects.