Impact, Where It Counts.
AI-Powered RFQ Assistant

Turn messy inquiries into
sales-ready RFQs — automatically

A visual walk-through of how SpiceWorx's AI assistant converts informal product inquiries from industrial distributors into structured, qualified quotation requests.

Try the Live Demo
System Flow

End-to-end process

From the customer's first message to two emails in the inbox — here's every decision point in the system.

Customer Inquiry 1. Product Discovery Natural language chat → catalog search Enough specs? AI evaluates context Yes No → ask follow-up 2. AI Qualification Specs · qty · voltage · urgency 3. Customer Requests Quote Fills contact form → submits 4. LLM Extraction Chat history → structured RFQ JSON Readiness Score > 70% complete? High Low Flag Missing Gaps → emails 5. Email Generation AWS SES dispatches both emails Customer Email Summary · gaps · Qs next steps Sales Brief RFQ JSON · score · products · transcript Sales Team Ready to Prepare Quotation
The Problem

Same inquiry — two completely different outcomes

Industrial distributors receive incomplete, unstructured inquiries every day. Each one requires multiple back-and-forth rounds before quoting can begin.

✗ Without AI
"Need a Schneider emergency stop button for a control panel."
  • Model/series unknown — sales must ask
  • Voltage & contact config missing
  • Quantity & delivery location unknown
  • Multiple round-trips delay quoting
✓ With AI
"I need a 22mm emergency stop. Schneider preferred, for a 220V panel."
  • Product category & brand extracted
  • Missing info flagged automatically
  • Smart follow-up questions generated
  • Structured RFQ ready for quoting
How It Works

5-step demo flow

From natural language inquiry to two structured emails — in under a minute.

1
Product Discovery
Customer describes their need in plain language. The AI searches the catalog and suggests matching products.
Natural language input
Catalog search
Brand suggestions
2
AI Qualification
The assistant conversationally gathers missing specifications — quantity, voltage, application, and urgency.
Specs & quantity
Voltage & current
Urgency & location
3
RFQ Handover
Customer clicks "Request Quotation." The existing SpiceWorx quotation form is reused — no new workflow needed.
Contact details captured
Existing form reused
Zero friction handoff
4
AI Extraction
The full chat history is sent to an LLM. It extracts structured RFQ data — including gaps, confidence, and a Sales Readiness Score.
LLM JSON extraction
Confidence score
Readiness score 0–100%
5
Email Generation
Two emails are auto-generated and sent via AWS SES — a confirmation to the customer and a sales brief to the team.
Customer confirmation
Internal sales brief
Sent via AWS SES
Data Model

What the AI extracts

Every conversation is parsed into a structured JSON payload with these fields. Highlighted fields are critical for quoting.

Product
Category Brand Model Specifications Quantity
Technical
Voltage Current Rating Application
Customer
Name Email Company Phone
Commercial
Urgency Delivery Location
AI Analysis
Confidence % Readiness Score Missing Info Follow-up Qs Product Suggestions

Highlighted fields are critical for quoting — the AI prioritizes these in follow-up questions.

Sales Readiness Score

How qualified is the inquiry?

Every RFQ gets a 0–100% completeness score so sales knows exactly what's ready to quote and what still needs follow-up.

95%
Ready

Best-case: all fields captured

95%
Quote-ready immediately
Product, specs, quantity, and delivery location all known. Sales can prepare the quotation now.
40%
Clarification required
Quantity, voltage, and delivery location are missing. AI generates targeted follow-up questions.
Email Outputs

Two emails, auto-generated

One confirmation goes to the customer. One structured brief goes to your sales team. Both sent via AWS SES, instantly.

Technical Architecture

Built on what's already there

The POC adds minimal new code — reusing the existing FastAPI backend, LLM integration, and AWS infrastructure.

Frontend
Demo Page New
  • —/en/demo-rfq-assistant.html
  • —Based on existing AI Knowledge Systems demo
  • —Chat UI + quotation form
  • —Real-time RFQ card preview
Backend
API Endpoint New
  • —POST /rfq/analyze
  • —Input: chat history + customer details
  • —Output: structured RFQ JSON
  • —Output: two email drafts
Reused Infrastructure
Existing Stack Reuse
  • —FastAPI backend
  • —Existing LLM integration
  • —AWS hosting + S3
  • —SES for email dispatch
  • —Existing quotation form
What's Being Sold

This is not a chatbot

Industrial distributors don't buy AI. They buy business outcomes. The AI is just the mechanism.

Faster Quotations
Structured RFQs arrive ready to quote — no back-and-forth to collect specs.
Better-Qualified Leads
Every inquiry arrives with specs, context, and a readiness score — no guessing.
Less Sales Effort
AI handles spec collection and follow-up questions so sales focuses on quoting.
Better Customer Experience
Customers get instant, structured confirmation — not silence after submitting a form.
Roadmap

Phased rollout

V1 is intentionally minimal — validate business value first, then expand channel and CRM integration in later phases.

Now — V1 POC
Core Demo
  • Product chat UI
  • RFQ handover form
  • Structured extraction
  • Customer email
  • Sales email brief
Phase 2
RAG + Catalog
  • Product catalog RAG
  • Technical manuals
  • Spec comparisons
Phase 3
Multi-Channel
  • Email channel
  • Embedded widget
  • White-label option
Phase 4
CRM & Analytics
  • Lead tracking
  • RFQ analytics
  • Sales performance
  • Inquiry trends

The success criterion

"Can this work with our product catalog?"

When a prospect asks that, the demo has succeeded. It's no longer a chatbot — it's a Sales Enablement System for industrial distributors.

Try the Live Demo
🌿
SpiceWorx AI Ask me anything about SpiceWorx