Flow Marketplace

Inspiration

Negotiating prices online is often time-consuming and inefficient. Both buyers and sellers spend hours messaging back and forth, manually monitoring price changes, and struggling to align on the right deal. We wanted to solve this by creating Flow Marketplace, an AI-driven marketplace where intelligent agents negotiate on behalf of both parties.

What it does

Flow Marketplace is an AI marketplace where buyers and sellers can delegate negotiations to intelligent agents.

• For Sellers: Upload your product, set your target price, absolute minimum price, and timeline. The AI agent will handle negotiations for you.
  
• For Buyers: Share your intent and desired price range, and let the AI agent negotiate across sellers to secure the best deal.
  
• AI Chatbot: Captures buying or selling intent in conversational form.
  
• Hero Scenario:
  
  • Seller side: Sellers specify price goals and preferences.
    
  • Buyer side: Buyers input their budget and desired product.
    
  • Both sides: AI agents negotiate automatically to find the best match.

How we built it

• Leveraged Plastic Labs’ Honcho for orchestration and agent-to-agent negotiation logic.
  
• Used Flowglad for workflow automation and session management.
  
• Designed session tagging for intent recognition (buy vs. sell).
  
• Integrated with modern authentication frameworks like Supabase Authentication.
  
• Built a chatbot-driven user experience to make onboarding seamless and natural.
  

Challenges we ran into

• Designing a fair negotiation mechanism that balances buyer and seller goals.
  
• Ensuring trust and transparency when users let AI negotiate on their behalf.
  
• Managing session tagging and intent classification to ensure smooth transactions.
  
• Integrating external authentication and marketplace workflows in a secure way.

Accomplishments that we're proud of

• Successfully created a working AI marketplace demo with autonomous negotiation.
  
• Built a conversational chatbot interface to capture user intent.
  
• Demonstrated how AI agents can save both buyers and sellers time while ensuring fairness.
  

What we learned

• How to design negotiation strategies that balance multiple constraints.
  
• The importance of session tagging and intent recognition in building conversational commerce.
  
• Authentication and user trust are critical for adoption.
  
• Hands-on experience with Honcho and Flowglad, and how they streamline AI-driven workflows.
  

What's next for Flow Marketplace

• Expand AI negotiation strategies to handle multi-party and bulk deals.
  
• Introduce rating and reputation systems for both buyers and sellers.
  
• Add support for more complex product categories beyond simple listings.
  
• Enhance explainability of AI negotiations so users can see why a deal was made.
  
• Explore partnerships with marketplaces to integrate AI-driven haggling at scale.

DEMO Use Case

• Problem: Negotiation is time-consuming; buyers and sellers waste hours messaging and monitoring price changes.
  
• Demo Flow:
  
  1. Seller uploads product and specifies target price, minimum acceptable price, and timeline.
     
  2. Buyer expresses interest with their budget and requirements.
     
  3. AI chatbot tags session as "buy" or "sell."
     
  4. AI agents negotiate automatically to find a fair deal.
     
  5. Both sides save time while reaching optimal agreements.
     

Built With

• next.js