Inspiration
Many small businesses run ads without any structured review of the creative itself. Most feedback is subjective, expensive, or comes after money has already been spent. We wanted a simple way to evaluate ad creatives early, using AI, so obvious design and messaging issues could be fixed before launch.
What it does
AI Ad Manager allows users to upload an ad screenshot and receive instant, structured feedback. Using Gemini AI, the system analyzes the creative and identifies its strengths, key issues, and specific improvement actions. This helps users understand what works, what doesn’t, and what should be changed to improve ad effectiveness.
How we built it
The frontend is built with React, providing a clean upload and results interface. The backend is implemented using Java Spring Boot, which handles file processing, AI requests, and data persistence. Ad screenshots are analyzed using the Gemini API, and results are returned in a structured JSON format. PostgreSQL is used for storage, and the application is deployed on Render.
Challenges we ran into
We faced issues with AI response consistency, rate limits, and handling large or malformed JSON outputs. Managing API quotas and ensuring reliable parsing of Gemini responses required careful prompt design and validation. Deployment and environment configuration also required troubleshooting to keep the live demo stable.
Accomplishments that we're proud of
We successfully built a fully working, live web application within the hackathon timeline. The system delivers clear, actionable feedback instead of generic AI text. We also achieved a smooth end-to-end flow—from upload to analysis—using Gemini’s multimodal capabilities.
What we learned
We gained hands-on experience with Gemini’s multimodal reasoning, prompt control, and response handling. We also learned how important structured outputs and validation are when integrating AI into real products, especially under real-world constraints like rate limits and deployment environments.
What's next for AI Ad Manager
We plan to add creative performance scoring, comparison between multiple ad versions, exportable reports, and user history tracking. Long-term, we aim to support multiple ad formats and platforms, making AI Ad Manager a practical tool for everyday marketing workflows.
Log in or sign up for Devpost to join the conversation.