• 

  Architecture: Eight AI agents coordinate supply-chain recovery through one central Coordinator.

• 

  Ideation: StockRescue turns supplier delays into actionable recovery plans through coordinated AI agents.

• 

  Research: Research into small-business challenges inspired smarter inventory and supplier decisions

• 

  Architecture: Eight AI agents coordinate inventory, supplier discovery, quote evaluation, and recovery planning.

• 

  Workflow: From one request to a recovery plan, with human approval before purchasing.

StockRescue: The AI Team That Rescues Your Business from Supply-Chain Disruptions

Eight specialized AI agents work together to help businesses investigate supply-chain disruptions, discover alternative suppliers, evaluate risks, and coordinate recovery plans while keeping purchasing decisions under human control.

💡 Inspiration

Small businesses often struggle with supplier delays, inventory shortages, changing prices, and missed delivery deadlines. Resolving these problems manually requires switching between spreadsheets, supplier directories, phone calls, and logistics updates.

We wanted to explore a better approach: what if a business owner could describe a disruption in one message and have a coordinated team of AI agents investigate the problem and prepare a recovery plan?

That idea became StockRescue — a multi-agent AI system designed to make supply-chain recovery more organized and accessible.

🚀 What StockRescue Does

A business owner can describe a problem such as:

"I need 30 sarees by Friday, but my current supplier is delayed. My budget is ₹2 lakh. Help me find alternatives and recover the order."

StockRescue coordinates specialized tasks to help investigate the shortage, identify affected orders, discover alternative suppliers, evaluate quotes, prepare procurement actions, and monitor delivery risks.

Instead of managing every step separately, the business owner gets a coordinated workflow for exploring recovery options, with human approval before purchases are committed.

🏗️ Architecture & Multi-Agent Orchestration

StockRescue uses a coordinator-led architecture built around ASI:One, ASI1 Mini, Agentverse, and the Agent Communication Protocol (ACP). The Coordinator organizes the business request and connects the work of specialist agents.

The architecture also incorporates persistent business data through Supabase, a payment-enabled premium-report workflow, supplier discovery, procurement approval, and delivery monitoring.

The workflow follows this path:

Business Request → Coordinator → Specialist Agents → Recovery Options → Human Approval

Each specialist handles a defined responsibility, while the Coordinator brings their contributions together into a unified recovery workflow.

🤖 Eight Agents on Agentverse

StockRescue separates supply-chain recovery into eight specialized agents.

Agent	Agentverse Handle	Function & Responsibility	Profile
StockRescue Coordinator	stockrescue	Conversational orchestrator, ASI1 Mini reasoning loop, interactive response cards, and human-in-the-loop enforcement.	View Agent
Inventory & BOM	sr-inventory	Calculates physical stock against committed balances and quantifies net shortages.	View Agent
Job Impact	sr-jobimpact	Assesses work-order criticality, uncovered unit demand, and deadline-breach risk.	View Agent
Supplier Discovery	sr-suppliers	Supports certified-supplier discovery, failed-vendor blacklisting, and OEM compatibility verification.	View Agent
Quote Risk & Planner	sr-quoterisk	Evaluates sourcing options using multiple criteria, landed-cost calculations, and single- or split-sourcing plans.	View Agent
Procurement	sr-procure	Prepares purchase-order drafts, supports cryptographic fingerprinting, and requires operator approval.	View Agent
Vendor Voice & Comms	sr-vendorcomms	Supports RFQ script generation, voice synthesis, and supplier communication through configured integrations.	View Agent
Delivery Watch	sr-delivery	Supports inbound tracking, urgent logistics monitoring, and schedule-variance alerts.	View Agent

🏁 How StockRescue Addresses the Fetch.ai Judging Criteria

Criterion	What StockRescue Brings
Functionality & Technical Implementation (25%)	A coordinator-led system with eight specialized agents covering inventory, job impact, supplier discovery, quote evaluation, procurement, vendor communication, and delivery monitoring.
Use of Fetch.ai Technology (20%)	Agentverse-hosted agent profiles, ASI:One and ASI1 Mini integration, ACP-based agent communication, and a payment-oriented premium-report workflow using Fetch.ai's Agent Payment Protocol.
Innovation & Creativity (20%)	A coordinated AI recovery team that connects shortage analysis, supplier alternatives, quote-risk evaluation, procurement preparation, and delivery monitoring in one workflow.
Real-World Impact & Usefulness (20%)	Helps small businesses investigate supply-chain disruptions, compare possible recovery options, and reduce fragmented manual coordination.
User Experience & Presentation (15%)	A conversational entry point, coordinated agent responsibilities, interactive responses, and human-reviewed purchasing actions.

These weights follow the judging criteria in the Homie example you shared. Confirm that they match the official criteria for your specific hackathon before submitting.

✨ What Makes StockRescue Different

• Multi-agent collaboration: Eight specialized agents contribute to one supply-chain recovery workflow.
• Risk-aware sourcing: Supplier discovery and quote evaluation help organize alternative sourcing options.
• Human-in-the-loop procurement: Purchase drafts and proposed actions remain subject to approval.
• Persistent business state: Supabase supports storing and retrieving business information.
• Payment-enabled services: The premium-report workflow explores payment-based access to additional business insights.
• Communication and delivery support: Dedicated agents address supplier communication and delivery monitoring.

🧠 Development Process & Team Collaboration

We approached StockRescue as a coordinated multi-agent system rather than a single chatbot. The work involved defining specialist responsibilities, organizing the central coordination flow, considering persistent business state, and connecting supplier-related workflows with procurement and delivery monitoring.

Our team contributed across backend and agent development, frontend and user experience, payment workflows, and pitching.

The goal was to make each agent's responsibility clear while keeping the overall experience centered on one business problem and a practical recovery plan.

🧩 Challenges We Faced

One of our main challenges was coordinating multiple specialist agents without losing the context of the original business request. Each agent has a different role, but its findings must contribute to a useful overall plan.

We also had to balance automation with business control. Recommendations can help a business owner evaluate alternatives, but purchasing actions need appropriate approval.

Integrating operational information, supplier options, quote risks, delivery constraints, and persistent business state into one coherent workflow was another important design challenge.

📚 What We Learned

StockRescue helped us explore how specialized AI agents can divide complex operational problems into manageable tasks and coordinate their outputs.

We learned that multi-agent systems require more than separate agents: they need clear responsibilities, coordinated information flow, relevant business context, and appropriate human oversight.

🔮 What's Next

• Improve supplier-data quality and sourcing evaluation.
• Strengthen quote-risk analysis and recovery planning.
• Expand delivery-status and logistics integrations.
• Improve workflow observability, reliability, and error handling.
• Refine payment-enabled premium reports.
• Make recovery recommendations easier for business owners to evaluate.

🌍 Our Vision

We envision StockRescue as a practical AI operations team for small businesses — helping owners investigate disruptions, explore alternatives, and coordinate recovery with less manual effort.

One business problem. Eight specialized agents. A coordinated path to recovery.

🔗 Try StockRescue

Resource	Link
GitHub Repository	View StockRescue source code
Agentverse	Open the individual agent profiles in the table above.

Built With

• agent-communication-protocol-(acp)
• agentverse
• asi1-mini
• asi:one
• elevenlabs
• fastapi
• fetch.ai
• langfuse
• postgresql
• python
• supabase
• twilio