Offomo – AI-Powered Deal Discovery

Inspiration

Online shopping should be simple, but finding the best deal often means searching dozens of websites for coupons, cashback offers, promo codes, flights, and discounts. We wanted to build an AI-powered platform where users could simply ask what they’re looking for in plain English and instantly receive the best available deals. Our vision is to make saving money as easy as having a conversation.

What it does

Offomo is an AI-powered deal discovery platform that helps users find the best products, promo codes, cashback offers, and travel deals across Canada and the United States. Instead of relying on keyword searches, users can ask natural questions like “Find me the best laptop under $1,000” or “Show me flights to Toronto next weekend.” Offomo understands intent, corrects spelling mistakes, expands synonyms, and returns relevant results in a conversational format.

How we built it

We built Offomo using Next.js, TypeScript, Supabase, Clerk, and Vercel, with OpenAI powering the AI experience. We use embeddings and semantic search to understand user intent, combined with structured search and business logic to surface the most relevant deals. The application also includes intelligent query classification, typo correction, session management, and a scalable architecture designed to continuously ingest new promotions and merchant offers.

Challenges we ran into

One of our biggest challenges was combining AI-powered semantic search with real-world shopping requirements. A highly relevant result isn’t always the best result if it’s expired, unpopular, or no longer available. We had to design a hybrid ranking system that balances semantic similarity with freshness, popularity, and business rules while keeping responses fast and accurate. Creating a conversational shopping experience that feels natural and trustworthy also required significant prompt engineering and retrieval optimization.

Accomplishments that we’re proud of

  • Built a production-ready AI-powered deal discovery platform from the ground up.
  • Created a conversational search experience that understands natural language instead of relying on keywords.
  • Developed a scalable search pipeline using semantic embeddings and hybrid retrieval.
  • Successfully launched Offomo on both iOS and Android, making AI-powered savings accessible across platforms.
  • Designed a modern, fast, and responsive user experience optimized for everyday shoppers.

What we learned

Building Offomo reinforced that great AI products require much more than integrating an LLM. Success depends on high-quality data, thoughtful retrieval systems, strong backend architecture, and continuous evaluation of search quality. We also gained valuable experience in semantic search, embeddings, hybrid retrieval, prompt engineering, authentication, deployment, and designing AI-first user experiences.

What’s next for Offomo

Our next goal is to make Offomo even more personalized and intelligent. We’re working on personalized recommendations, real-time price tracking, deal alerts, expanded travel search, additional merchant integrations, and more advanced hybrid search ranking. As we continue to grow, our mission remains the same: help people save time and money by making deal discovery as simple as asking AI.

Built With

  • ai
  • application
  • coupon
  • deals
  • discount
  • flight
  • hotels
  • mobile
  • next.js
  • rag
  • shopping
  • supabase
  • travel
  • vercel
Share this project:

Updates