CRM 2.0 System Design

1. System Overview & Design Philosophy

1.1 System Positioning

This system is a CRM 2.0 Sales Management Platform built on the NocoBase no-code platform. The core goal is:

Let salespeople focus on building customer relationships,
not data entry and repetitive analysis.

The system automates routine tasks through workflows and leverages AI to assist with lead scoring, opportunity analysis, and more — helping sales teams work more efficiently.

1.2 Design Philosophy

Principle 1: Complete Sales Funnel

End-to-end sales flow:

Why design it this way?

Principle 2: Configurable Sales Pipeline

Different industries can customize pipeline stages without modifying code.

Principle 3: Modular Design

• Core modules (Customers + Opportunities) are required; all others are optional
• Disabling a module requires no code changes — configure via the NocoBase admin UI
• Each module is independently designed to minimize coupling

2. Module Architecture & Customization

2.1 Module Overview

The CRM system uses a modular architecture — each module can be independently enabled or disabled based on business needs.

2.2 Module Dependencies

2.3 Pre-configured Editions

2.4 Module–Table Mapping

Core Module Tables (Always Required)

2.5 How to Disable a Module

Simply hide the module's menu entry in the NocoBase admin panel. No code changes or table deletions required.

3. Core Entities & Data Model

3.1 Entity Relationship Overview

3.2 Core Table Details

3.2.1 Leads Table (nb_crm_leads)

Lead management with a simplified 4-stage workflow.

Stage flow:

New → Working → Qualified → Converted (Customer/Opportunity)
        ↓            ↓
   Unqualified   Unqualified

Key fields:

3.2.2 Customers Table (nb_crm_customers)

Customer/company management with foreign trade support.

Key fields:

3.2.3 Opportunities Table (nb_crm_opportunities)

Sales opportunity management with configurable pipeline stages.

Key fields:

3.2.4 Quotations Table (nb_crm_quotations)

Quotation management with multi-currency and approval workflow support.

Status flow:

Draft → Pending Approval → Approved → Sent → Accepted / Rejected / Expired
              ↓
          Rejected → Revise → Draft

Key fields:

3.2.5 Orders Table (nb_crm_orders)

Order management with payment tracking.

Key fields:

3.3 Table Summary

CRM Business Tables

Base Data Tables (Shared Module)

3.4 Supporting Tables

3.4.1 Comments Table (nb_crm_comments)

General-purpose comment/note table, linkable to multiple business objects.

3.4.2 Customer Shares Table (nb_crm_customer_shares)

Enables multi-user collaboration and permission sharing on customers.

3.4.3 Opportunity Collaborators Table (nb_crm_opportunity_users)

Supports sales team collaboration on opportunities.

3.4.4 Regions Table (nb_cbo_regions)

Country/region base data dictionary.

4. Lead Lifecycle

Lead management uses a simplified 4-stage workflow. When a new lead is created, a workflow can automatically trigger AI scoring to help sales quickly identify high-quality leads.

4.1 Status Definitions

4.2 Status Flow

4.3 Lead Conversion Flow

The conversion UI presents three options simultaneously; users can create or link:

• Customer: Create a new customer or link to an existing one
• Contact: Create a new contact (linked to the customer)
• Opportunity: Must create an opportunity

Fields recorded after conversion:

• converted_customer_id: Linked customer ID
• converted_contact_id: Linked contact ID
• converted_opportunity_id: Created opportunity ID

5. Opportunity Lifecycle

Opportunity management uses configurable pipeline stages. When a stage changes, a workflow can automatically trigger AI win probability prediction to help sales identify risks and opportunities.

5.1 Configurable Stages

Stages are stored in nb_crm_opportunity_stages and can be customized:

5.2 Pipeline Flow

5.3 Stagnation Detection

Opportunities with no activity will be flagged:

-- Calculate stagnation days
UPDATE nb_crm_opportunities
SET stagnant_days = EXTRACT(DAY FROM NOW() - last_activity_at)
WHERE stage NOT IN ('won', 'lost');

5.4 Won / Lost Handling

When Won:

1. Update stage to 'won'
2. Record actual close date
3. Update customer status to 'active'
4. Trigger order creation (if quotation was accepted)

When Lost:

1. Update stage to 'lost'
2. Record loss reason
3. Record competitor ID (if lost to a competitor)
4. Notify manager

6. Quotation Lifecycle

6.1 Status Definitions

6.2 Approval Rules (To Be Refined)

Approval flow is triggered based on the following conditions:

6.3 Multi-Currency Support

Design Rationale

USD is used as the unified base currency for all reports and analysis. Each monetary record stores:

• Original currency and amount (what the customer sees)
• Exchange rate at the time of transaction
• USD equivalent (for internal comparison)

Currency Dictionary (nb_cbo_currencies)

Currency configuration uses a shared base data table for dynamic management. The current_rate field stores the current exchange rate, synced by a scheduled task from the latest record in nb_crm_exchange_rates.

Exchange Rate History (nb_crm_exchange_rates)

Records historical exchange rate data. A scheduled task syncs the latest rate to nb_cbo_currencies.current_rate.

Note: Quotations link to nb_cbo_currencies via currency_id FK and read the rate directly from current_rate. Opportunities and orders use a currency VARCHAR field for the currency code.

Monetary Field Pattern

Tables with monetary amounts follow this pattern:

Applied to:

• nb_crm_opportunities.amount → amount_usd
• nb_crm_quotations.total_amount → total_amount_usd

Workflow Integration

Exchange rate fetch logic:

1. Business operations read rate directly from nb_cbo_currencies.current_rate
2. USD transactions: rate = 1.0, no lookup needed
3. current_rate is synced by scheduled task from the latest nb_crm_exchange_rates record

6.4 Version Management

When a quotation is rejected or expires, it can be copied as a new version:

QT-20260119-001 v1 → Rejected
QT-20260119-001 v2 → Sent
QT-20260119-001 v3 → Accepted

7. Order Lifecycle

7.1 Order Overview

Orders are created when a quotation is accepted, representing a confirmed business commitment.

7.2 Order Status Definitions

7.3 Order Data Model

nb_crm_orders

nb_crm_order_items

7.4 Payment Tracking

nb_crm_payments

8. Customer Lifecycle

8.1 Customer Overview

Customers are created upon lead conversion or opportunity win. The system tracks the full lifecycle from acquisition to advocacy.

8.2 Customer Status Definitions

8.3 Customer Health Score

Health score is calculated from multiple factors:

Health score thresholds:

if (health_score >= 90) status = 'advocate';
else if (health_score >= 70) status = 'active';
else if (health_score >= 50) status = 'growing';
else status = 'at_risk';

8.4 Customer Segmentation

Automatic Segmentation

9. Email Integration

9.1 Overview

NocoBase provides a built-in email integration plugin supporting Gmail and Outlook. Once emails are synced, workflows can automatically trigger AI analysis of email sentiment and intent, helping sales quickly understand customer attitudes.

9.2 Email Sync

Supported mailboxes:

• Gmail (via OAuth 2.0)
• Outlook / Microsoft 365 (via OAuth 2.0)

Sync behavior:

• Bidirectional sync for sent and received emails
• Automatic association to CRM records (leads, contacts, opportunities)
• Attachments stored in the NocoBase file system

9.3 Email–CRM Association (To Be Refined)

9.4 Email Templates

Sales can use pre-built templates:

10. AI Capabilities

10.1 AI Employee Team

The CRM integrates the NocoBase AI plugin, using the following built-in AI employees with CRM-specific tasks configured:

10.2 AI Task List

AI capabilities are divided into two independent categories:

1. AI Employees (Frontend — User-Triggered)

Users interact directly with AI employees via frontend blocks to get analysis and recommendations.

2. Workflow LLM Nodes (Backend — Auto-Executed)

LLM nodes embedded in workflows, triggered automatically via table events, action events, or scheduled tasks — independent of AI employees.

Note: Workflow LLM nodes use prompts with schema-defined output to produce structured JSON, which is then parsed and written to business data fields — no user interaction required.

10.3 AI Fields in the Database

11. Workflow Engine

11.1 Implemented Workflows

11.2 Planned Workflows

12. Menu & Interface Design

12.1 Admin Menu Structure

12.2 Dashboard Views

Sales Manager View

Sales Rep View

Executive View

Document Version: v2.0 | Last Updated: 2026-02-06