top of page

Rishma Mendhekar

gxrl talk

gxrl talk is a prototype for a womxn's health AI chatbot app.

I created gxrl talk for my Ammerman Center Senior Integrative Project. 

Cover Screen.png
Chat - message input.png
Settings - Stats – 1.png

The opening screen, chat page, and calendar page of gxrl talk.

gxrl talk consists of three main sections: a chat page, a calendar, and a settings page where users keep track of profile information. The chat page is where the most important functionality takes place: this is where users can ask Cynthia-- the app's AI chatbot-- any health questions. Cynthia can also remind users about appointments and to take medications. The calendar page displays logs created by the users; users can track anything from their period, to their medications, to gym days. The settings page allows users to manage their personal information and calendar logs. Artificial intelligence was not included in the prototype completed during the year.

I created gxrl talk over the course of two semesters during my senior year at Connecticut College for my senior integrative project at the Ammerman Center for Arts and Technology. I used Adobe XD for prototyping and Android Studio to code a prototype of the app.


For an informational video and a full walkthrough of the app, please click here. The password is gxrltalk.


I came up with basic idea of gxrl talk based on my own experience with reproductive health and hearing the stories of other women on community-based health apps. Coming from a conservative family, I never had anyone show me the ropes, as they say. Fast forward to college: I downloaded a women's health app which featured a community section; users would post their health questions and receive answers from other members. For example, someone asked how long after starting birth control they would be fully protected. The downside-- someone answered one month, someone else said three months, and my doctor had told me one week. It became clear that crowd-sourcing health information wasn't effective. Additionally, I had studied established artificial intelligence systems related to health (Florence and Woebot, amongst others) in previous courses.

It was then that I developed the idea for gxrl talk-- what was really needed was an accessible but reliable source of women's health information. After discussion with my SIP adviser, I conducted further research on women's health and health education, which only proved my hypothesis. The statistics were worrisome:


In 2016, Women between the ages of 15 and 19 had the highest rate of gonorrhea and the second highest rate of chlamydia of any age group in the United States.

1, 2

24 states require public schools to teach sex education. Only 20 require the information to be medically accurate, and 27 states recieve “abstinence education” funding.



Initial sketch for the calendar page.

From there I created my initial sketches (left) and built them out into Adobe XD, along with interactions. After creating a color palette, I applied it to the app and sketched out the first version of Cynthia, the face of the chatbot. 


I met with my adviser and Ammerman peers weekly to get feedback on initial designs. This was essentially my process for the entire first semester: add to the design, get critique, and modify the design once again.

From the get-go, I knew that the app would have three main parts: the chat page, the calendar, and the settings page. My adviser and I agreed that the overall tone of the app would be friendly and conversational since the point of the app was to be accessible. We also agreed that the chatbot would need to have a "face"-- the more human the chatbot seemed, the more users would trust it. In a nutshell, our main design principles were: trustworthy, friendly, and knowledgable.

The biggest design push happened midway through the semester. The design still looked flat, and I still wasn't done with the settings page. After a month of redesigns, at the end of the semester-- 4 different color palettes, and a slow journey into gradients-- I landed on what was essentially the final design.

Chat - sent message 1 copy.png
Calendar - month copy.png
Chat - message input copy.png
Settings - Stats copy.png

The three main pages (home, calendar, settings), October 2017.

Three main pages, December 2017.


Coding the prototype began in second semester. I had never used Android Studio before,

so the semester began with simple tutorials to get the hang of the IDE and the language. Each page took about a week to complete, and there was one final redesign of the bottom navigation bar. Although coding the app was my main priority, the majority of my time spent working on gxrl talk was for supporting materials (3 posters and a video) that I would display in Conn's senior art minor show in April.

I created three informational posters and a video to showcase the app and my inspiration behind it. The show was up for two weeks, but I was able to use the materials for subsequent presentations as the end of the semester approached. I finished coding the prototype in late April for a second showcase hosted by the Ammerman Center.



In the future, I would like to integrate the AI component into gxrl talk.

bottom of page