Inspiration
People often avoid traditional surveys because they get nothing in return, which makes them hard to promote and easy to ignore. We realized the real problem is not the questions themselves, but the context in which they are asked. In games like Imposter or Spyfall, players answer very similar questions willingly and often truthfully, because their goal is to blend in and not be suspected. This insight inspired us to move questions out of forms and into gameplay, where answering feels natural, engaging, and purposeful.
What it does
SurveyDonkey transforms classical surveys into question and text based games. Instead of filling out forms, users play short games like Imposter and Top Tier lists. While they play, their choices and answers are continuously translated into structured survey data. The platform supports different question types and collects insights in real time without disrupting the user experience.
How we built it
We designed the platform around two main layers. The first layer is the game layer, which handles gameplay logic, user interactions, and how questions are presented inside games like Imposter and Top Tier.
Between these two layers, we integrated AI to bridge gameplay and data. Using the Gemini API, AI helps transform classical survey questions into game friendly prompts, interpret player responses, and map in game actions back into meaningful survey data.
On the backend, we integrated a PostgreSQL database to store responses, game states, and aggregated insights. The database is hosted and managed as part of our infrastructure, allowing us to support continuous data collection, real time updates, and scalable growth as more users play.
Challenges we ran into
How we built it: We used React and TypeScript to build the website, focusing on a modern, fast, and highly gamified user interface. On the backend, we integrated a PostgreSQL database to store survey data, gameplay states, and aggregated insights, along with a light mode, professional survey creation and analytics page that allows teams to analyze data and extract actionable insights.
To power the intelligence of the platform, we used Google’s state of the art Gemini 3 Flash model to transform traditional survey questions into engaging game mechanics. The entire application is hosted on a Hetzner instance, giving us full control over performance, scalability, and operating costs.
Accomplishments that we're proud of
We are proud of finding an innovative idea with strong real world potential and validating that it can work in practice. Even more, we successfully designed and implemented a functional end to end solution in less than two days, proving both the feasibility of the concept and the speed at which it can be built.
What we learned
We learned that collecting good feedback does not require long or intrusive forms. Well designed, lightweight interactions can be more accessible, engaging, and sometimes even more informative. We also learned how critical prompt design and clear constraints are when using AI to reliably transform gameplay interactions into structured, high quality data.
What's next for SurveyDonkey
Next, we plan to scale the platform by adding more text based games such as Truth or Dare. We will start collecting additional signals like response time and subtle player interactions to unlock deeper behavioral insights. We also plan to further leverage AI to analyze this richer data and automatically generate reports with clear, actionable insights. On the product side, we aim to let users define context such as favorite brands, workplace, or interests, enabling better targeting and more relevant feedback from the right audience.
Log in or sign up for Devpost to join the conversation.