Sorcara Copilot
PROJECT OVERVIEW
Sorcara Copilot is an AI-powered soucing agent that empowers retailers and wholesalers to navigate the complexity of strategic and custom product sourcing.
I led the end-to-end design process to create an intuitive AI copilot experience that simplifies complex business processes, making intelligent automation accessible to businesses while providing clear, actionable insights.
📍Problem
The B2B global sourcing market is booming, valued at approximately $18.6 trillion in 2023 and projected to grow at a CAGR of around 18.2% through 2030. However, this lucrative market presents significant challenges for retailers and wholesalers, including:
• Overwhelming Strategic Decision Complexities:
The sourcing journey requires businesses to navigate a complex landscape of decisions, from analyzing customer trends and designing specifications to verifying suppliers, negotiating terms, managing logistics, and adapting to fluctuating markets.
• Supplier Network Constraints:
Limited supplier networks further challenge businesses by hindering their ability to find suitable partners, resulting in inflated costs, production delays, and slower time-to-market.
📍Existing Solutions
Traditional sourcing methods rely on outdated catalogs, demanding extensive manual research, or human assistants limited by their under-digitalized networks and knowledge, leading to a loss of control over supplier relationships and potential hidden commissions.
RESEARCH & DISCOVERY
Target Users
Retailers & Wholesalers
Small to medium businesses seeking efficient product sourcing solutions, ranging from dropshippers to established retailers across apparel, electronics, and consumer goods industries.
Research Method
User interviews & Competitive analysis
Conducted 25+ user interviews with retailers and wholesalers, analyzed existing sourcing platforms (Alibaba, Global Sources), and studied B2B procurement workflows.
Key Findings
🤖Phase 1 Smart RFQ (request for quotation) Agent
Our Smart RFQ Agent translates everyday language into industry-specific terminology. When users input vague descriptions like "cat scratcher," our AI system automatically generates precise specifications—surface materials, dimensions, compliance requirements—reducing sourcing time from weeks to days.
⚠️Limitation
We did several user interviews and discovered that when making purchase decisions, users still struggle with not being able to see actual products and must rely on submitting requirements and waiting for supplier responses. This lack of visual context creates uncertainty and delays in decision-making.
🔍Phase 2 Sourcing Assistant
The agent is built on six core components:
Product & Supplier Database:
Accesses a massive network of over 500M+ products, 45K suppliers, and 60+ logistics partners.
Real-Time Market Database:
Utilizes dynamic data on sales, trends, and more to stay ahead of market shifts.
Sourcing Knowledge Graph:
Leverages a proprietary knowledge graph to understand the relationships between product specifications, categories, consumer trends, and compliance requirements.
Query Analysis:
Accurately interprets natural language input to understand user needs and diversify workflow.
Workflow Planning:
Breaks down complex sourcing goals into smaller, manageable tasks.
Dynamic Response Delivery:
Constructs tailored responses by leveraging a library of modular templates to address specific user needs and queries.
🔑Key Use Cases
The agent allows retailers and wholesalers to freely discuss their sourcing needs—whether their language is vague or precise, professional or naive—with a system that understands, processes, and infers user intentions, intelligently directing tasks to the appropriate workflow.
Key use cases include:
Find white label products
Find suppliers for highly customizable products
Generate RFQ & Tech Packs
Access sourcing research and consultation
Compare quotes, Negotiate Terms
Connect with fulfillment options and status
01 Find White Label Products
Buyer Queries Example: "A pet tracker for small dogs (2kg - 5kg)"
02 Find Suppliers for Highly Customizable Products
Buyer Queries Example: "I want to start a print-on-demand business for custom pet hoodies. I need a supplier who can print custom designs on hoodies for dogs of various sizes. I'm looking for high-quality, soft materials that are comfortable for pets."
03 Sourcing Research & Consultation
Buyer Queries Example:
- • "What are the regulations for using the term 'organic' on pet food labels in the US?"
- • "Help me research the most trendy features to source pet food"
- • "What is your recommendation on specifications for best selling pet clothes in winter?"
🔥Achievements
In November 2024, Sorcara MVP was launched, empowering 50+ retail brands and dropshippers across 6 industries, ranging from apparel to consumer electronics. Customers reported a staggering 35% reduction in sourcing costs compared to their previous reliance on Alibaba or human-operated sourcing companies, coupled with a remarkable 75% faster time to market. This improvement translated to thousands - millions in increased profit, demonstrating the undeniable impact of AI-powered sourcing.
To further validate Sorcara MVP's capabilities, we conducted a 2-hour evaluation session with 15 customers, comparing our agent to competitors like Accio (Alibaba's buyer agent) and Shopping with Perplexity AI. This evaluation encompassed a wide range of tasks, including: query understanding, product and supplier discovery, research and analysis, and document / negotiation generation.
Across all these tasks, Sorcara MVP achieved an average total score of 7.6/10—a 25% improvement over competitors—and maintaining a less than 10% bad case rate, based on customer evaluations.