Smarter supply chains for small & medium factories.
SupplyLink is a smart platform that uses machine learning to forecast demand, keep inventory at the right level and plan transport — giving SME manufacturers the supply chain intelligence that used to be reserved for large corporations.
SMEs power the economy — but plan their supply chains on guesswork.
Large manufacturers run teams of analysts and expensive planning software. Small and medium factories make the same decisions with spreadsheets and instinct.
of all businesses worldwide are small and medium-sized enterprises.
of global employment comes from SMEs.
Source: World Bank, Small and Medium Enterprises (SMEs) Finance
Uncertain demand
Guessing next month's orders leads to overproduction or missed sales.
Inventory imbalance
Too much stock ties up cash; too little stops the production line.
Costly transport
Unplanned, half-empty deliveries push up operating costs.
Three engines. One supply chain.
SupplyLink applies machine learning to the three decisions that matter most in an SME supply chain.
Demand forecasting
Predict what customers will order, so production follows real demand.
- Forecasts from historical demand
- Earlier warning of shifts
- Production planning support
Inventory optimization
Keep raw materials and finished goods at the right level — no more, no less.
- Fewer stock-outs
- Less cash tied up in stock
- Smarter reorder timing
Transport optimization
Plan shipments that use resources efficiently and reduce operating costs.
- Better use of vehicles
- Lower delivery costs
- More dependable lead times
From raw data to confident decisions.
- 01
Connect data
Bring together demand records, inventory levels and transportation data.
- 02
Train models
Regression, neural network and ensemble models learn the patterns in your operations.
- 03
Optimize
Optimization methods tune each model for better accuracy and lower cost.
- 04
Decide
Clear recommendations on what to produce, stock and ship — aligned with your goals.
Three families of machine learning, each with a job.
SupplyLink's approach compares and combines established model types, choosing the right tool for each supply chain problem.
Regression models
Capture how demand, stock and costs move together — fast, transparent and easy to explain.
Neural networks
Learn non-linear patterns across many variables that simpler models miss.
Ensemble models
Combine many models into one stronger, more reliable prediction.
Machine learning beats traditional planning.
- F/01Better predictionMachine learning models outperformed traditional strategies in forecasting.
- F/02Better resource useResources were used more efficiently across the supply chain.
- F/03Lower operating costsOptimization methods improved model performance and reduced operating costs.
- F/04Aligned operationsDecisions became better aligned with supply chain objectives: speed, quality and dependability.
Data the models learn from
Source: Alkubaisi et al., IEEE ISMSIT 2024. doi:10.1109/ISMSIT63511.2024.10757215
The road ahead.
The directions the research identifies for the next generation of supply chain intelligence.
ML optimization core
Regression, neural network and ensemble models for demand, inventory and transport.
Hybrid methods
Combining model types to get the strengths of each in a single system.
Real-time data
Live data feeds so recommendations update as conditions change.
Sustainability KPIs
Tracking the environmental performance of the supply chain alongside cost.
Business know-how meets machine learning.
SupplyLink was founded in 2024 by a research team uniting supply chain and business expertise with computer science.
Administrative & Financial Sciences
Defines the business problems — costs, inventory policy, logistics and the KPIs that matter to factory owners.
Computer Science
Faculty of Science. Builds the machine learning models, the optimization engine and the platform itself.
Frequently asked questions
What is SupplyLink?
A smart logistics platform that uses machine learning to optimize the supply chains of small and medium-sized factories — forecasting demand, managing inventory and planning transport.
Which machine learning models does it use?
Regression models, neural networks and ensemble models, improved further with optimization methods.
What data does it need?
Demand records, inventory levels and transportation data — the information most factories already collect.
Is machine learning really better than traditional planning?
The research behind SupplyLink found that machine learning outperformed traditional strategies in prediction and resource utilization, and that optimization reduced operating costs.
Why focus on small and medium factories?
SMEs make up around 90% of businesses and over half of global employment, yet rarely have access to advanced supply chain tools.
Backed by peer-reviewed research.
Read the IEEE study behind SupplyLink's machine learning approach to supply chain optimization.
Alkubaisi, S., Fenjan, A., Al-Sharify, T. A., Sadiq, S., Furaijl, H. B., & Ahmed, S. R. (2024). Machine learning models for optimizing supply chain management. In 2024 8th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT) (pp. 1–9). IEEE.
https://doi.org/10.1109/ISMSIT63511.2024.10757215