Customer 360 and ecommerce operations in one connected workspace
A self-hosted Customer 360 and ecommerce CRM designed to unify customer identity, orders, communications, support activity, segmentation, and connected business systems in one operational workspace.
Portfolio imagery is approved for public presentation; private customer data, credentials, endpoints, and infrastructure details remain excluded.
System overview
Customer information is often fragmented across transactions, authentication, email, accounting, support, and analytics tools. Staff can see individual events while still lacking a complete view of the customer relationship.
Thumbmark CRM is the customer intelligence and operational CRM layer of the DoneDealion ecommerce ecosystem. It gives internal teams a joined view of customers, orders, communications, support cases, segmentation, analytics, and connected-system activity.
The platform centralizes context that would otherwise remain distributed across WooCommerce, customer identity services, transactional email, accounting, support workflows, and AI services—without replacing the systems that remain authoritative for commerce, identity, or accounting.
Customer 360
Product focus
Joined customer and operational context
Internal teams
Primary users
Customer service, operations, and management
Self-hosted
Deployment
Containerized web application
Production system
Status
Continuing product development
The business problem
Ecommerce operations create useful customer signals in many separate systems. Without a coordination layer, identity, purchases, messages, complaints, and customer value must be reconstructed manually whenever staff need context.
The solution
Thumbmark creates a Customer 360 workspace where staff can understand who the customer is, what they purchased, which communications they received, what support activity exists, how their lifecycle segment was computed, and which connected system supplied each record.
Core features
Profiles, orders, addresses, communications, activity, cases, notes, tags, and segmentation in one customer record.
Structured case intake, assignment, evidence, replies, SLA monitoring, escalation, resolution, and public tracking.
Customer, order, segment, care, and communication reporting with honest unavailable-data states.
Code-computed Recency, Frequency, and Monetary groups for customer lifecycle and retention decisions.
An English- and Tagalog-capable assistant using permitted database tools for grounded CRM answers.
Connected-system scope, synchronization history, and separate connection, health, and activity signals.
Inside the platform
Large product views show how Thumbmark brings customer context, operations, and connected systems into one staff workspace.
Customer 360
Customer 360 brings identity and contact context together with commerce, communications, activity, support history, and connected-system provenance.
One staff-facing record for the customer relationship.
Customer Analytics
Dedicated analytics surfaces help staff interpret customer, order, segment, care, and communication patterns while keeping accounting authority with FinFlow.
Operational analysis beyond the dashboard summary.
Customer Segmentation
Application code and database facts compute Recency, Frequency, and Monetary scores. AI can explain aggregate results, but it does not calculate the segment itself.
Deterministic lifecycle groups with aggregate AI interpretation.
Customer Care
Customer Care connects complaints back to Customer 360 and gives staff structured tools for ownership, investigation, communication, SLA, escalation, and resolution context.
A traceable case workflow connecting support activity to Customer 360.
AI Assistant
Mika uses OpenAI function calling and database-backed tools to help authenticated staff ask natural-language questions in English or Tagalog.
Tool-backed conversational access to permitted CRM facts.
Integration Hub
Thumbmark presents the customer and operational picture while connected systems retain authority for commerce, identity, accounting, communication delivery, files, and AI inference.
Connection, health, and activity are represented as separate signals.
Staff Management
Staff administration combines invitations, account activation, roles, session revocation, and audit records for accountable internal access.
Role-aware administration and staff access control.
System architecture
Architecture groups are data-driven and intentionally avoid unverified production details.
WooCommerce remains the commerce and order authority.
Customer Ecosystem is represented as the customer identity and profile authority.
The Customer 360, operations, analytics, and coordination layer.
Bounded services support communication, accounting context, evidence, and AI.
Technology stack
Next.js 14.2
App Router staff workspace
React 18 + TypeScript
Typed interface implementation
Tailwind CSS + Radix UI
Responsive components and interaction patterns
Recharts
Operational data visualization
Python + FastAPI
Asynchronous API and business services
SQLAlchemy 2 async
Database models and queries
Pydantic 2
Request and response contracts
Alembic
Schema migrations
PostgreSQL 16
Customer, care, staff, audit, and telemetry persistence
Asyncpg
Asynchronous PostgreSQL connectivity
Docker Compose
Self-hosted service deployment
Nginx + HTTPS
Reverse proxy and secure delivery
OpenAI
Tool-backed assistance and fact-constrained explanations
Cloudinary
Customer Care evidence storage
My responsibilities
Challenges and solutions
Challenge
Commerce, customer identity, communication delivery, accounting, support, and AI each have different owners and data boundaries.
Solution
Thumbmark acts as a provenance-aware coordination layer while WooCommerce, Customer Ecosystem, and FinFlow retain authority for their domains.
Result: Staff gain a joined operational view without treating the CRM as the universal source of truth.
Challenge
CRM questions and charts require explanation, but model-generated calculations could conflict with stored business data.
Solution
SQL and application services compute metrics and RFM segments; AI receives bounded facts to explain, summarize, categorize, or assist.
Result: AI remains advisory while deterministic system state stays authoritative.
Development journey
Current foundation · completed
Customer 360, analytics, segmentation, Customer Care, Mika, staff controls, audit, and integration-health foundations are implemented.
Maturing · current
Identity reconciliation, behavior collection, engagement tracking, synchronization conflicts, and provider monitoring continue to mature.
Future work · planned
Returns, voucher lifecycle, stronger complaint-resolution architecture, reverse logistics, and future FinFlow/AHMS boundaries remain planned work.
System relationships
Commerce authority
WooCommerce customers, guest checkout context, orders, items, statuses, and totals flow into Thumbmark.
Live · inbound synchronization
Identity and profile authority
Global Customer ID, customer profile, contact, and address updates are represented through inbound service contracts.
Implemented · partial producer verification
Accounting authority
Linked accounting-related communication records appear in customer history; Thumbmark does not replace the financial ledger.
Live · narrow scope
Email delivery
Sender-owned recipient, type, source, delivery status, and sent-time records feed communication history.
Live · inbound activity
AI inference
Supports Mika, AI Explain, aggregate segmentation insight, and advisory Customer Care analysis.
Live · request and response
Evidence storage
Stores Customer Care attachment files while Thumbmark retains their metadata and case relationship.
Live · bounded file scope
Next conversation
Explore the wider portfolio or get in touch about an engineering opportunity or business system.