Retail & Inventory · LIOUS project
CSD Analytics & Forecasting
Powered by LIOUS AI
Developed by LIOUS Advisory Services, CSD Analytics combines canteen operations, inventory intelligence and machine-learning demand forecasting with LIOUS AI, an assistant powered by locally hosted Ollama models. The platform runs on organisation-controlled local infrastructure.
Project scope
- Custom Software Development
- Predictive Analytics
- Local AI Integration
The challenge
Canteen teams need to coordinate sales, stock, demand planning and reconciliation across multiple operational records. Historical reports alone make it difficult to anticipate replenishment needs and investigate exceptions.
- Bring sales, stock and procurement information into one workspace.
- Identify slow-moving items, stock shortages and supply variations.
- Make demand forecasts understandable to operational teams.
- Run the application and AI services within local infrastructure.
What we built
LIOUS developed a locally deployed web application that connects operational registers, analytics and forecasting. LIOUS AI helps users explore operational information and interpret forecasts while managers retain responsibility for purchasing and stock decisions.
- Central dashboards for sales, stock valuation, inventory movement and forecast records.
- Python-based statistical and machine-learning demand forecasting.
- ABC inventory classification and fast-, slow- and non-moving analysis.
- Demand, invoice and goods receipt workflows with supply variance tracking.
- Expiry monitoring, replenishment alerts and sales reconciliation.
- Role-based access, historical data imports and Excel/PDF reporting.
Built for local deployment
The application runs on organisation-controlled local infrastructure, with Docker services for the web application, operational data, forecasting and Ollama-based AI inference. This is a private installation; no public application login is provided.
Platform capabilities
LIOUS AI assistant
Conversational support for exploring sales, inventory and demand indicators, using compatible models hosted locally through Ollama.
Forecasting and reorder planning
Python analytics combine historical sales and seasonal patterns to support projected demand, stock requirements and replenishment review.
Inventory intelligence
ABC classification, configurable thresholds and movement analysis help teams prioritise high-value stock and identify understocked or overstocked items.
Procurement and receipt tracking
Demand registers, invoice records and goods receipts support comparisons between demanded, invoiced and received quantities, including short, excess and damaged stock.
Expiry and reconciliation
Near-expiry monitoring, stock-loss reporting and sales-versus-collection comparisons help teams investigate operational exceptions.
Imports and management reporting
Validated operational data imports, historical analysis and Excel/PDF exports bring sales, stock, demand and accounting information together.
How it works
- 01
Import operational data
Bring sales, stock and demand records into the platform with validation.
- 02
Analyse inventory
Review stock levels, product movement, consumption value and sales trends.
- 03
Forecast demand
Use historical patterns and seasonal signals to estimate replenishment requirements.
- 04
Review with LIOUS AI
Ask questions and interpret analytical outputs before making operational decisions.
- 05
Plan and reconcile
Track demand, invoices and received quantities; investigate supply and collection variances.
- 06
Monitor and report
Review reorder and expiry alerts and export operational reports.
Forecasts and AI explanations support human review. They do not guarantee future demand or replace authorised procurement decisions.
Operational value
A unified local workspace for inventory planning, operational reporting and AI-assisted analysis. The platform is designed to improve visibility, reduce fragmented recordkeeping and support informed replenishment decisions.
Technology
- React
- Node.js
- Express
- Python
- MySQL
- Docker
- Ollama