Docimate

Inspiration

The idea for Docimate started at home. Affan's dad runs a UPS store and spent hours each month sorting through messy invoices, forms, and contracts. Watching him stay late, struggling with disorganized paperwork, inspired us to create a solution that could automate the process and help small business owners like him.

What it does

Docimate is an AI-powered document assistant that extracts key information from unstructured files like PDFs and reports, then organizes it into clear, searchable tables. It allows users to ask questions directly to the AI about their documents and create custom templates to extract data in specific formats or packages — making document handling faster, easier, and error-free.

How we built it

We built Docimate using Google Gemini for natural language processing and intelligent text extraction. The AI analyzes unstructured documents and outputs structured data in JSON, which makes it easy to transform into multiple export formats like CSV and PDF. We utilized Vercel AI SDK to allow seamless integration between the frontend and the AI engine (mainly with structuring data). The frontend was designed with simplicity in mind so that anyone can use it, even without technical experience.

Challenges we ran into

One major challenge was crafting the right prompts to maintain accuracy across different document types. Scanned PDFs, handwritten forms, and inconsistent layouts often confused the model, so we had to carefully design and iterate on our prompts to guide the AI in recognizing and extracting key fields consistently.

Accomplishments that we're proud of

We’re proud that Docimate can now analyze and simplify full documents in seconds — something that used to take hours. Seeing Affan's dad use it to handle his own store paperwork efficiently was the most rewarding moment for our team. Additionally, winning Best Overall, Best Finance, and Best UI at WinHacks 2025 was a huge validation of our work and impact.

What we learned

We learned how to combine AI, design thinking, and real-world problem solving to create something truly useful. We not only deepened our understanding of NLP, text extraction, and user experience design, but also learned the importance of building tools that solve real, human problems.

What's next for Docimate

We plan to expand Docimate's capabilities by integrating multi-language support, cloud-based document storage, and team collaboration features. Our goal is to make Docimate a go-to productivity tool for small businesses, educators, and professionals worldwide, helping them turn messy documents into meaningful, structured data effortlessly.