BidOrNot

AI-powered public tender sourcing for IT companies operating in Europe. Find Finnish public tenders your team can actually win.

Comment

Inspiration

I preferred a SaaS bidding platform because it can demonstrate the full potential of using AWS databases. A large amount of data is fetched from an API and stored alongside user data and AI-generated data.

What it does

This SaaS platform helps companies find the best option to win public tenders. While the idea isn't entirely new, in 2026 we can significantly reduce the time needed to find suitable tenders with the help of AI. An LLM analyzes a company's profile and matches it with relevant procurements. The LLM can then prepare a plan to win the procurement. The results can be discussed with the company's team to support decision-making.

How I built it

  • Week 1 – brainstorming, ideation, exploring Hilma's API service, retrieving raw data, analyzing how to use the data, creating JSON schemas and database schemas, making the first request to AI for data analysis, writing system prompts, creating a design system for the whole project, building the main screen using an ASCII-style approach, and organizing all data into a single structure later used for AI-driven web development.
  • Weeks 2–3 – implementation from scratch. Used all $35 credits from v0.app, plus other agentic AI tools (GitHub Copilot, Antigravity, Zed, Kiro, Augment). Connected to an AWS Aurora database.

Challenges I ran into

Working with Next.js, AWS databases, v0.app's agentic workflow, and Vercel as a deployment platform were all brand new to me.

Accomplishments I'm proud of

With this new agentic workflow, I can ship a full-stack app in weeks—not months, as was the case with my first full-stack project in 2023.

What I learned

  • Predicting all project details in advance is crucial. It's better to spend more time on preparation than to spend time on rebuilding later.
  • I should have used an AWS database from the start. Initially, I developed the project with data stored in local storage and IndexedDB, which cost me 1.5 days to migrate to AWS.

What's next for BidOrNot

  • Test the product with real potential customers to see how well this SaaS fits their needs.
  • Write tests for the current functionality and check for the top 10 OWASP security issues—I skipped these to speed up the project.

Built With

Share this project:

Updates

Submission history