Google Gemini Tutor/Classroom Conversations: Helping students and teachers 24/7


The rise of technology in the classroom is undeniable, and has only accelerated after the COVID-19 pandemic. The Google Suite of products emerged from this time as a dominant tool in classroom management and learning with its Google Classroom platform. Forbes reports that in 2020 46% of US learners were entirely online, and the trend has continued in higher education with 60% of college students enrolled in distance learning courses in 2021.

As of February 2021, Google reports that more than 150 million students, educators and school leaders use Google Classroom to teach and learn, a 40 million user increase over the previous year.

Educators represent diverse subject matter, ages, experience with technology, attitudes towards distance learning, and administrative support. Students too, represent a wide variety of learning abilities, language abilities, ages, communication styles, support needs, and tech literacy.

Is there a way to help both educators and students outside of traditional classrooms get quick support for a variety of learning needs?
Potential Users
Potential users are extremely varied. Users include administrators, educators, students, and parents. Users include minor and adult learners, in traditional and non-traditional classrooms. They include native and non-native speakers of a classroom language.
Research objective: We want to know pain points people experience staying connected and getting support in distance learning classrooms.

Product objective: We want to make a product that provides useful and quick support for educators and learners who use Google Classroom as a learning platform.


I combined primary research via interviews of potential users with competitor research of similar products. My goal was to understand what online tools educators and students are currently using to improve learning outcomes, what frustrations people experience with their current tools, what unique learner populations are not being currently served, and what variety in use of online tools occurs between in-person, hybrid, and full remote learning.

Primary Research: Interviews

I spoke with 5 participants; 3 primarily as educators, 1 primarily as a student, and 1 as both student and educator. All had experience with Google Classroom as an LMS, and used other online tools to supplement their learning/teaching.

Affinity Mapping

Key Themes

I found 7 themes:
  • Students like a way to quickly, and preferably by chat/text, communicate with each other and the teacher
  • Students like to ask questions to each other, and often prefer it to asking the teacher
  • The teacher is sometimes hard to reach in remote/distance learning contexts, sometimes due to not having notifications turned on
  • There is often a resistance among teachers to learning new tech solutions, either due to time needed to learn new tool, resistance to changing preferred methods, or concerns about technical issues that will arise
  • Distance learning poses challenges for teachers, students, and parents, and all need support, which is often not accessible
  • A wide variety of tools have been used: Google Classroom, Google Drive (without Classroom), Blackboard, Brightspace, Zoom, Quizlet, Whatsapp, Discord, Time4Learning, Teams, YouTube
  • The decision of what tool to use is sometimes determined by the administration and sometimes by the teacher

Key Takeaways

  • Some products offer live support, but this requires paid accounts
  • Cost of these platforms is hard to access, making it difficult for some institutions/educators to know if they will be able to budget the cost, making free preferrable
  • Many offer real-time text communication, which users mentioned liking
  • Accessibility options could be improved for some LMS

Creating User Personas

User Personas

Two personas were created to represent the student and teacher perspective of using Google Classroom in a higher education context.

Jan is an adult international student at a US university in a hybrid course, and as such he sometimes struggles to understand the course content. He wants to ask for help, but it is difficult to get access to the teacher or students outside of class time. He wishes he could easily communicate with his class or ask clarifying questions at any time.

Grace is a professor who had to transition to online teaching after the pandemic, and is being pushed to use tools that she isn't confident with. She wishes she could get live support or had an assistant who could take some of the work load off of her adapting her materials to the online space. She often has questions at inconvenient times, and the administrative support isn't working with her schedule.

Developing Solutions

Once I had a good idea of the motivations, goals, and pain points of potential users, I started to ideate on how to help solve the problems they face through Point of View Statements and turning them into How Might We...? questions.

Priority Features

At this point, I was realizing AI might be able to assist in solving these questions. Google has access to so many tools that aren't integrated into Classroom yet, and bringing in Google Gemini (previously named "Bard") AI, Google Translate, and Google Meet, I might be able to get students and teachers the help they need.

I used user stories to brainstorm possible features for a new Google Classroom feature. From this, I ranked them into priority levels, high, medium, and low. This created our list of must-have features for adding a new feature to the LMS.

Priority Features

  • A place to type questions
  • A place to view answers to questions
  • A way to translate answers into any supported language
  • A way to start a chat with students and teachers
  • A way to export answers to Google Docs or another text editor
  • A place to ask tech support questions

Low Fidelity Wireframes

Operating within a pre-existing product was challenging, since I didn't want to negatively impact usability of the current product. Where would be the most logical place for students to interact with a conversation board? In additions, how do users feel towards AI bots? Were they likely to expect an AI assistant to be where "chat bots" often live, in the lower corner? Finally, Gemini AI is still under development, and as such, Google is making changes to the icons and interaction patterns constantly. Was the current iteration of interaction with Gemini AI understandable by users?

Mid Fidelity Wireframes

After sharing with colleagues, design elements following Google's Material Design 3 were added.

Testing Mid Fidelity Wireframes

At this stage, I tested the usability of the product with 5 participants in a moderated usability test. We tested 4 tasks: Asking Gemini AI Tutor a question, Exporting that answer to Google Docs, Translating the answer, and Posting a question on the conversation board.

Key Insights

  • Users had difficulty completing the task to ask the AI tutor a question; they defaulted to asking a person for help
  • Users responded positively to the concept of an AI tutor, indicating high likelihood of use
  • Users liked how the AI tutor was with them constantly, and immediately accessible
  • The translate function was easy to complete and rated very useful
  • Users wanted the other interaction icons with Gemini Tutor labeled due to uncertainty in meaning
  • Users liked the conversation channels, but mentioned reservations due to shyness and embarrassment
  • "People" tab confused users, thinking they would message people from there
  • Overall, they like the simplicity and function of the two features

High Fidelity Wireframes

I used Google's Material Design 3 system to add styling to the pages.

Testing and Revising

For this iteration, I did an unmoderated usability test on to try to see how users interacted with the site without conscious or unconscious guiding from me as a facilitator. I tested the same 4 tasks as in the previous test: Asking a question to Gemini AI Tutor, Exporting the answer to a Google Doc, Translating the answer, and Posting a question in the conversation boards.

Users liked the product, rating very high on likelihood of use. They especially liked how they could ask questions to the AI Tutor without "appearing dumb", a recurring theme. In contrast, users still expressed uncertainty about the conversation channels, due to a fear of being thought of as "dumb". Saving face with real people is something important to consider in this design. Users were still misdirected by the "People" tab, which offers students no interaction, only a way to view all students in the class. This can be added into the conversations tab if it is important information.


  • Change AI Tutor button to a circle to draw more attention, and add animation
  • Remove "People" tab from student view, as it is directing attention away from conversations but students are not able to interact with anything on this tab

Key Revised Screen

Product walk-through

Interact with the prototype in Figma


In this video, I give a tour of the 4 flows tested during product development: Asking Bard Tutor for help, Exporting the answers to Docs, Translating the answer, and Posting a question on the Conversations board.