AI & AUTOMATION
Building systems that connect
commerce, data and operations.
I use automation, APIs and lightweight software to remove repetitive work, connect fragmented processes and make business decisions easier to scale.
The goal is not automation for its own sake. It is better visibility, fewer manual hand-offs and systems that people can actually use.
Python · SQL · APIs · Google Apps Script · Looker Studio · Generative AI
SYSTEMS I’VE BUILT
Three systems built to solve real operating problems.
ECOMMERCE WORKFLOW AUTOMATION
Automating the operating layer behind ecommerce.
Problem
Scaling ecommerce created repeated manual hand-offs across order receipt, payment verification, inventory confirmation, processing, fulfillment, dispatch and reconciliation.
What I built
Mapped the complete operating workflow and progressively automated repetitive steps using the Shopify API and Python. The system connected ecommerce activity with the operational processes required to fulfil orders reliably rather than treating the website as a standalone sales channel.
System design
- Order receipt and processing
- Payment verification
- Inventory confirmation
- Fulfillment workflow
- Dispatch coordination
- Reconciliation and reporting
- API-driven workflow automation
Outcome
Within the first year, virtually the entire order-to-dispatch workflow was automated, with physical packaging remaining the primary manual step.
PythonShopify APIWorkflow AutomationEcommerce Operations
Related case study → Revlon Sri Lanka
ECOMMERCE WORKFLOW AUTOMATION
APPLICATION & DATABASE SYSTEM
Turning import-export operations into a structured SQL-backed system.
Problem
Commercial information was distributed across order folders, purchase orders, proforma invoices, commercial invoices, packing lists, payment records and shipping documents. Finding the complete status of an order required navigating multiple files rather than querying a structured source of truth.
What I built
Built a local trading-operations application with a Python application layer and a relational SQL database backend using SQLite. The database stores structured commercial data; the application is the working interface for retrieving, linking and managing order, product, company and document information.
System design
- Relational SQL data model
- Structured order records
- Product-line data
- Company records
- Commercial-document relationships
- Payment and transaction data
- Search and data retrieval
- Document linking
- Excel export workflows
PythonSQLSQLiteData ModelingDocument ProcessingLocal ApplicationExcel Automation
APPLICATION & DATABASE SYSTEM
REPORTING AUTOMATION
Turning recurring outlet sales reporting into a repeatable system.
Problem
Outlet sales reporting required repeated manual consolidation across locations and reporting periods, making analysis time-consuming and increasing the risk of inconsistent data.
What I built
Built an automated sales-reporting workflow using Google Sheets and Google Apps Script to standardize data capture, calculations and recurring reporting across outlets.
Business use
Created a consistent view of outlet-level sales performance, making it easier to compare locations, identify sales trends and analyze performance over time while reducing repetitive manual processing.
Google SheetsGoogle Apps ScriptSales ReportingWorkflow Automation
REPORTING AUTOMATION
SELECTED SYSTEMS
Supporting analysis and reporting.
GLOBAL · MARKETING INTELLIGENCE
Bringing global campaign performance into one decision view.
LG Business Solutions campaigns operated across Asia, Europe, the United States and the Middle East.
Built a unified Looker Studio reporting environment to compare country-level demand, lead delivery, investment and performance more consistently across markets.
Looker StudioGoogle AdsGoogle AnalyticsMulti-market Reporting
Related case study → LG Business Solutions
ECOMMERCE · DATA-DRIVEN OPERATIONS
Connecting ecommerce demand with product availability.
Used sales history, demand patterns and commercial data to support forecasting and inventory planning rather than treating marketing demand and stock availability as separate problems.
The workflow helped inform decisions around sales planning, product availability and inventory requirements.
Decision SupportSales & inventory planning
Sales DataDemand ForecastingInventory PlanningEcommerce Analytics
AI IN THE WORKFLOW
AI is a component of the system, not the system itself.
I use generative AI where it can reduce low-value cognitive work: research synthesis, structured drafting, content development, information organization and workflow acceleration.
AI output remains an intermediate layer. Business rules, commercial context and human judgment determine the final decision or output.
Research Synthesis · Structured Drafting · Content Workflows · Process Design · Information Organization · Workflow Experimentation
HOW I THINK ABOUT AUTOMATION
Automate the friction.
Keep the judgment.
01Start with the business problem
Automation is useful only when it improves something that matters: speed, accuracy, visibility, consistency or scale.
02Build around the source of truth
A good system should make data ownership and information flow clearer — not create another disconnected layer.
03Keep humans in the decision loop
Forecasting, campaign optimization and generative AI all require commercial and market context that systems alone cannot fully interpret.
04Make the output usable
The best system is not the most technically complicated one. It is the one people can use consistently.
Technology should solve a business problem.
I build systems where commerce, operations and data intersect — using the simplest technology capable of making the process faster, clearer and more scalable.